'P V fet., /..
iiS^^r»
r^ ‘"Çr
.xO
--
lîPÉ^If 4 M?:''' '‘; ■■f'
S^fi^WÂÏ^i 'N’K
■■ ■ W’ ' :è^ ^: •"., '• ■ ' ’■ M
'V'
■'
■ m.^i .
)
Annals
OF THE
Royal Botanic Gardens,
Peradeniya.
t
EDITED BY
J. C WILLIS, M.A., F.L.S.,
DIRECTOR.
VOLUME I.
dune, 1SÛ1— Sepîemliâry 1S02,
(EültîmiîXî :
UEORdE J, A, SKEEN. GOYERNMENT PRINTER, CEYLON.
Ktmtfm :
DULAU & CO,, 37, SOHO SQUARE, LONDON, W.
\ Ail rights of Reprodvction and Translation reserved f
s 8 0 • 7 s
ANNALS OF ÏHE ROYAL BOTANIC GARDENS.
PERADENIYA.
VOL I., l!»0l-02.
Dates of Pill) lication of Parts,
Part I. pp. 1-2G
Part II. pp. 27-180
Part III. pp. 181-260
Part IV. pp. 267-466
June 27, 1901
December 22, 1901
May 31, 1902
December 9, 1902
Part IV. dated September, 1902, was printed and ready for publi-
cation early in that month, but was delayed by late arrival of some of
the plates which left Europe in August.
Dates given at the ends of certain papers indicate the days upon
which they were handed to the Editor complete ; no alterations were
made afterwards.
Instructions to Binder,
Plates XXIII. and XXVII. are omitted. Plates IV.-XXXVIII.
to be placed in one block at the end of the volume, and to be cut
down to size of text.
Supplements to be removed before binding and retained separately
until complete.
( iv )
(H)NTENTS.
Departmental and Editorial Notes : page
Opening of tlie Experiment Station at Gangariiwa, Pera-
deniya ... ... ... 264
Personal Notes ... ... ... Î80, 266
Publications ... ... ... 266
The New Branch Garden at Nuwara Eliya ... 265
The New Laboratory at Hakgala Garden ... 179
The New Resthoiise at Peradeniya ... ... 265
Original Papers and Notes :
Gardinee, J. S., see Willis.
Holtermann, C., Fungus Cultures in the Tropics (wit-h
plate I.) ... ... ... 27
Michelin et Cie., see Willis.
Willis, J. C., A Revision of the Podostemaceæ of India
and Ceylon ... ... ... 181
- — ~ Material for demonstration of developing Embryos
(Klugia) ... ... " ... 179
MM. Michelin et Cie. ’s Analyses of Ceylon India-
rubbers prepared by Mr. Parkin ... ... 253
*— Note on the Flora of Minikoi ... ... 39
=— Note on the Product of the Latex of the Jak (Arto-
carpus integrifolia') ... ... ... 258
Notes of Indian Travel ... ... 259
— — - Studies in the Morphology and Ecology of the Po-
dostemaceæ of Ceylon and India (with plates IV.-
XXXVIII.) ... ... ... 267
The Royal Botanic Gardens of Ceylon and their
History ... ... ... 1
The Royal Botanic Gardens of Ceylon as a Centre
for Botanical Study and Research ... ... 17
Willis, J. C., and Gardiner, J. S., The Botany of the
Maidive Islands (with plate II.) ... ... 45
Wright, Herbert, Observations on Dracaena reflexa
(with plate III.) ... ... ... 165
Pith for Microscopists ... ... 258
PA(iF.
( V )
Personalities and Historical Notes :
Bamber, M. K., 13 ; Bird, G., 6 ; Cameron, W., 7 ; Carruthers,
J. B., 13 ; Clark, P. D. G., 10 ; De Alwis, Alfred, 266, Harama-
nis 6, William 6, 266; Farmer, J.B., 12; Fraser, G., 7 ; Gardner.
G., 7 ; Giesenhagen, K., 252 ; Green, E. E., 13, 180 ; Hartog,
M. M., 9 ; Hooker, Sir J. D., 8 ; Jonville, J.,3 ; Keeble, F. W,,
12 ; Kerr, W,, 3 ; Lear, J.G.,6 ; Macmillan, H. F., 12 ; Macrae,
J., 6 ; Moon, A., 3 ; Morris, Sir D., 9, 10 ; Nock, W., 10 ; Nor-
mansell, ÏÏ. T., 6 ; Ondaatje, W. C., 6 ; Parkin, J., 13 ; Potter.
M. C., 12 ; Preyer, A., 251 ; Stuhlmann, F., 251 ; Thwaites,
G. H. K., 7 ; Trimen, H., 10 ; Walker, A., 6 ; VVard, H. M.,
10 ; Watson, J.G., 6 : Wilcox, E. M., 251 : Willis. J. O.. 12 ;
Wright, Herbert, 13.
Reviews«
Giesenhagen, A Monograph of Niphobolus
Nestler, A Simple Test for Thein
Parkin, The Coagulation of Latex
Preyer, The Agriculture of Southern Asia
Stuhlmann, The Settlement of European Peasantry in
the Mountains of the Tropics
Tropical Agricultural Journals
Ward, Disease in Plants
Wilcox, Glimpses of Tropical Agriculture
Zimmermann, The Bacterial Warts on the Leaves of
Pamtta indica and other Rubiaceæ ...
252
252
252
251
251
175
173
251
252
BRIEF SUBJECT INDEX.
Only those orders, genera, and species are mentioned here about
which important notes are made ; species enumerated in lists of
Floras, &c., are not usually given here.
New genera or species described for the first time are given in
italics ; genera or species re-defined or transferred to other groupings
are given in spaced type.
PAGE
Anatomy of Dracæna reflexa ... ... 165
Artocarpus integrifolia. Latex-product ... ... 258
Bacterial Warts on Leaves ... ... ... 252
Birds, Seed-dispersal by ... ... ...140,450
Botanical Research in Ceylon ... ... 17
( Vi )
PAGE
Botanic Gardens, Ceylon, History, &c.
1, 17
Ceylon, Flora of
... 181, 267
Chagos Islands, Flora of
51
Correlation (in Evolution) ...
434
Currents, Seed Dispersal by,..
145
D i c r æ a
... 216, 340
— - dichotoma
... 220, 349
elongata
... 219, 340
— — minor
... 222, 362
— styl osa
... 22.5, 355
W a 1 1 i c h i i
... 22.3, .351
Disease in Plants
173
Dispersal of Seeds, &c.
... 137, 450
Distribution, Geographical ...
...39, 45,
137, 159, 448, 450
Dorsiventrality
434
Dracæn a reflexa
165
Drugs, Maldivian
134
Dyes, Maldivian
133
Ecology of Podostemaceæ ...
267
Embryo Development in Klugia
179
Endogeny ...
415
Evolution, Mechanism of, &c.
434
Exogeny
415
Farm er ia
... 246, 396
indica, n. sp.
... 248, 403
■ metzgerioides
... 247, 397
Fibres, Maldivian
133
Flower, Morphology of the ...
... 194, 426
Fruit, Morphology, &c.
... 195, 426
Foodstuffs, Maldivian
135
Fungi
9,27
Fungus Culture
27
Germination
167, 267, &c.
Griffitheila
... 231,364
H o 0 k e r i a n a
... 233,364
Growth in Thickness (Dracæua)
165
Hemileia vastatrix
32
Hydrobryum
... 235, 375
Griffithii
... 238, 394
Johnson ii
... 241, 393
lichenoides
... 242,375
olivaceum
... 239, 379
n. sp.
... 239, 379
India, Notes of Travel in ...
259
— Flora of (Podostemaceæ)
... 181,267
Tndiarubber, Analyses of, &c.
253
( 'Pii )
PAGE
Island Floras (competition of plants,
order of
appearance,
&c.)
• ••
155
Klugia (embryogeny)
179
Laboratory Notes : flask for nutrient solutions
34, fungus-
culture 27, material for embryo development 179, pith for
microscopists 258.
Laccadive Islands, Botany of
...
45
Lawia
... 209,306
zeylanica
... 213, .307
Maidive Islands, Flora of (list, origin
of flora.
local names,
vegetation of atolls, economic products, &c.)
45
Man, Seed Dispersal by
137
Minikoi, Flora of
• » •
,39
Morphology of Podostemaceæ
... 181, 267
Niphobolus ceylanicus
252
Oceanic Islands, Floras of ...
159
Oils, Maldivian
133
Ornamental Plants, Maldivian
136
Pavetta, Bacterial Warts in (review)
252
Perfumes, Maldivian
133
Phylogeny
... 452,454
Pith for Microscopists
258
Plant Associations
155
Plant Names, Maldivian
126
Podostemaceæ
... 181,267
Podostemon
... 228,327
— — Barberi, n. sp.
. • .
... 230,336
- — subulatus
... 229,328
Polymorphism
... .365,408
Root Morphology
410
Rubiaceæ, Bacterial Warts in (review)
252
Seed Dispersal
137
Shoot Morphology
... 192,406
Sterilization
33
Tanning Stuffs, Maldivian ...
133
Thallus in Podostemaceæ ...
... 406,190
Thein, Test for
252
Timbers, Maldivian
. • .
136
Tristicha
... 207,290
— ramosissima
... 208,293
Tropical Agricultural Journals
175
Willisia
... 233,369
— — selaginoides
...
... 2.35, .370
—— sp. nov. ?
235
Wind , Seed Dispersal by
144
The Royal Botanic Gardens of Ceylon,
and their History.
BY
J. C. WILLIS.
rr^HE commencement of a new century, and with it of the
present Journal, affords a good opportunity for a brief
sketch of the present position and past history of the well-
known scientific establishment over which the writer has
the honour to preside. In recent years, especially, a con-
siderable expansion of the scope of the Department, and of
the work carried on by it, has taken place. In many ways
the history of the gradual enlargement of the establishment
reflects the general history of the nineteenth century in
Botany and its allied sciences and arts. In the early years
of the past century, when Botany consisted only of the study
of the external characters of plants, their classification, and
the investigation and cataloguing of the floras of the different
regions of the world, the Royal Botanic Gardens of Ceylon
were occupied principally with the collection and descrip-
tion of the wild plants of the Island. Towards the middle
of the century began the rise of Economic Botany, and the
then Director did splendid service in the introduction and
acclimatization of numerous useful and valuable plants from
other parts o| the world, whilst not neglecting the study of
the still very incompletely known vegetation of Ceylon
itself. Later still, and with the rise of Vegetable Pathology,
the fact began to be recognized, partly no doubt as the
[Annals of the Royal Botanic Gardens, Peradeniya, Vol. I., Pt. I., June, 1901.
49-01
(1)
2
WILLIS : HISTORY OF THE
result of the collapse of the great coffee industry by the
attack of the leaf disease, that science can afford services
’of as much value in aiding established industries to combat
disease as in introducing new industries, and this has
gradually led to the appointment of two officers upon the
staff of the gardens whose primary duty is the investigation
of the insects and fungi of the Island, especially those
likely to be injurious, and the discovery, if possible, of
means of combating their attacks. For the work of these
officers, and for other scientific work carried on here, a new
laboratory has been opened. Whilst the collection of the
local flora and the introduction and acclimatization of
plants from abroad is still vigorously carried on, these are
no longer the sole duties undertaken by the Department.
The condition of the various local industries is investigated
by the scientific staff, and experiments are carried on in
order to determine in what ways the methods or operations
of these industries may be profitably improved ; other
experiments are devoted to the discovery, if possible, of new
industries based on the cultivation of native or introduced
plants.
It will thus be seen that there is scarcely any branch of
modern botanical or agricultural science which is entirely
unrepresented in the present organization of the Department.
The general basis of the work is rigidly scientific, but the
production of results of immediate practical value to local
industries is steadily kept in view. Every facility for the
carrying out of researches in pure or applied Botany is so
far as possible afforded, both to workers on the regular staff
of the gardens and to workers from Europe and abroad, for
whom space is reserved in the laboratory.
It will be convenient to deal with the subject by the
historical method. The present headquarters of the Royal
Botanic Gardens, a Department of the Public Service of
Ceylon, are at Peradeniya, but this was by no means always
the case. The Dutch had a garden in Slave Island»
CEYLON BOTANIC GARDENS.
Colombo, but after the English conquest this was neglected
and sold by the Government. The first English Governor,
the Hon. F. North (afterwards Lord Guilford) had a small
private garden at Peliyagoda near Colombo, under the
superintendence of Joseph Jonville or Joinville, whom he
brought out as ‘‘ Clerk for Natural History and Agricul-
ture.” In 1800 Jonville accompanied General MacDowall’s
Embassy to Kandy, and made a collection of plants which
is now in the British Museum. He also drew some of the
plates in Cordiner’s “ Description of Ceylon.” Several
exchanges were made between the garden in his charge and
the gardens of the East India Company at Calcutta.
In 1810 Sir Joseph Banks, then President of the Royal
Society, was instrumental in causing the opening of the
first English Botanic Garden in Ceylon, under the super-
intendence of W. Kerr, who was transferred from Canton,
arriving in 1812. Seven acres of land were opened in Slave
Island, where the site is still indicated by Kew road, and
Kerr was placed in charge of this and of the garden at
King’s House as “ Resident Superintendent and Chief
Gardener.” Kerr brought with him several plants from
China ; his name is commemorated in the well-known
shrub Kerria japonica, introduced by him to Europe.
In 1813 the garden was moved to Kalutara, on the south-
west coast, the Colombo site having been found too subject
to flooding. The Government had resumed possession of an
unsuccessful sugar estate of 600 acres at Ugalboda, on the
left bank of the river, and upon this the garden was
reopened. In the following year Kerr died, and was suc-
ceeded by Alexander Moon, who arrived in Ceylon in 1817.
Under him the gardens were much improved, and in 1821,
six years after the conquest of the Kandyan kingdom, were
transferred to their present site at Peradeniya, four miles
from the centre of Kandy, on the Colombo road. The site
chosen was in most ways excellent, suffering chiefly in
the lack of a sufficient water supply, the river which almost
4
WILLIS : HISTORY OF THE
encircles it being at too low a level to be available for
watering the garden. The site is almost horseshoe-shaped,
occupying a peninsula round which flows the broad and
rapid stream of the Mahaweli-ganga, the principal river of
Ceylon. It is about one mile in length from north to south,
and has an area of 143 acres, ^prettily undulated. The
southern end is rocky, but with fairly good soil in many
parts ; the northern end consists largely of old river deposits
of sand and gravel. On the farther side of the river rise
the hills of Gangaroowa estate, one of which, to the north-
east, is about 700 feet higher than the garden and acts as an
efficient barrier against the force of the north-east monsoon.
The general level of the garden is about 1,550 feet above the
sea, and the climate is warm and damp, but very much
pleasanter than the climate of the wet zone of the low-
country, as exemplified in Colombo. The air has a bracing
freshness that is lacking in the plains, and the nights are
always cool. The climate is sufficiently warm to render
possible the cultivation of all but a very few tropical plants,
for which the nights are too cold in February and March
and the days too dry, and sufficiently cool to allow of the
cultivation of many sub-tropical plants suitable to medium
elevations in the mountain zone, but not capable of cultiva-
tion in the low-country. It is also sufficiently cold to allow
of energetic office or laboratory work with as little discom-
fort as in Europe. The successful Avorking of the Botanical
Department is in no small degree due to the choice by Moon
of so excellent a site as regards climate and position.
Before returning to the history, it may be well to say a
little about the climate of Peradeniya. The weather of the
year depends upon the two monsoons, in each of which there
is usually a very rainy commencing period, and a drier
termination ; the dry season of the north-east monsoon is
the more pronounced, and is generally known here as
the “dry” or “hot” season. The following table gives
the rainfalls, number of rainy days, approximate mean
CEYLON BOTANIC CARDENS.
5
temperatures, and approximate ranges of temperature for
each month of the year : —
Month.
Rain-
fall.
Inches.
Days.
Mean
Temp.
F.
Mean
Max.
F.
Mean
Min.
F.
Wind from
January
2-83
6
74
83
65
N.E.
February
P57
4
76
87
65
N.E.
March
4-67
8
77
88
66
N.E.
April
9-66
14
78
88
68
N.E-S.W.
May
716
12
78
86
70
S.W.
June
10-17
20
75
80
70
s.w.
July
6-84
17
75
80
70
S.W.
August
5-96
16
76
83
69
s.w.
September
6-40
15
76
85
67
s.w.
October
13-18
19
75
84
66
S.W-N.E.
November
9-74
16
74
84
64
N.E.
December
8-07
13
74
84
64
N.E.
Year ...
86-25
160
76
84
68
—
The figures for the temperatures are merely rough ap-
proximations, as no regular records have been kept at
Peradeniya. The highest recorded temperatures in the year
rarely reach 90°, the lowest 60°. The highest rainfall in
twenty-four hours seldom exceeds three inches, though
almost treble this amount has been recorded.
The rainfall of the year is well distributed, but there is a
somewhat severe drought in the first three months of the
year, sufficient to make the lawns look rather brown, and to
prevent that loading of the trees with epiphytes which is
seen in the wetter parts of the Island, and in such a garden
as Buitenzorg, where there is more than twice the rainfall.
Moon transferred to Peradeniya all the plants which could
be moved from Kalutara, and laid out the south-eastern part
of the ground, planting especially coffee and cinnamon. He
spent much time in the investigation of the flora of the
Colony, and collected largely in the Western Province, as
well as near Kandy and in Uva. In 1824 he published at
Colombo, in English and Sinhalese, his “ Catalogue of the
6
WILLIS : HISTORY OF THE
Indigenous and Exotic Plants growing in Ceylon.” Of the
1,127 plants native to the Colony there enumerated, 164 are
new species described for the first time. He made a consi-
derable herbarium, now mostly at Kew, and commenced the
library of the Department. In 1818 Harmanis de Alwis
Seneviratne was appointed writer under him, and displayed
such talent for drawing that Moon had him taught at his
own cost and appointed as draughtsman in 1823. This was
the beginning of the splendid series of coloured drawings
of the Ceylon flora and of other plants cultivated in the
gardens, which has been steadily continued to the present
time by H. de Alwis and his two sons, one of whom, William
de Alwis Seneviratne, Muhandiram, is at present the
draughtsman of the Department. In May, 1825, Moon died
of fever, and for some time the gardens were in charge of
Andrew Walker as Acting Superintendent. He was suc-
ceeded in 1827 by James Macrae, who collected many plants,
especially orchids From this time till 1844 the Department
languished, being chiefly used as a Government market
garden, the produce of which was sold in Kandy. Macrae
died in 1830, and after another interregnum under G. Bird,
James George Watson was appointed Superintendent in
1832. He collected plants at Jaffna and Puttalam and died
in July, 1838. The rubber avenue at the entrance to the
gardens was laid out by him in 1833, Another period of
interregnum under J, G. Lear," who came to Ceylon in 1837
or earlier as collector for Messrs. Knight of Chelsea, followed,
lasting till the appointment in 1840 of H. T. Normansell,
“ a clever young surgeon.” Lear collected and described a
number of orchids, and was one of the earliest tea planters ;
he planted tea at Nuwara Eliya in 1837. He laid out the
beautiful group of palms at the entrance of the gardens.
In January, 1843, Normansell died, and W. C. Ondaatje
acted as Superintendent till the arrival in May, 1844, of
* Many letters of his on garden matters are reprinted in “ Literary
Register,” 1890.
CEYLON BOTANIC GARDENS.
7
George Gardner, F.L.S., the well-known Brazilian traveller,
who was appointed on the recommendation of Sir William
Hooker. With his arrival the Department started on a new
career. Only one quarter of the land was then in use ; the
remainder was largely cleared and planted, new roads opened,
and many new plants introduced from other countries. A
new bungalow for the Superintendent was built in the centre
of the garden, and the entrance lodge erected. Gardner
travelled almost all over the Colony, and made large collec-
tions of native plants, many of which were new to science.
Unfortunately for the Department and for science, he died
of apoplexy at Nuwara Eliya, at the early age of 37, in March,
1849. A cenotaph, containing a brass to his memory, was
erected in the gardens in 1855. His herbarium was pur-
chased by the British Museum, and thus unfortunately lost
to the Colony.
From March to December, 1849, Mr. G. Fraser was Acting
Superintendent of the Gardens pending the arrival of the
newly-appointed Superintendent, George Henry Kendrick
Thwaites, who was for 31 years to control the destinies of
the Department, arid to do so much for scientific botany and
for the planting industries of Ceylon. Born in Bristol in
1812, Thwaites was 37 years old when he arrived in Ceylon,
and had already/ won a distinguished reputation in Botany.
Until about 1857 his duties were largely of a purely
scientific kind, but afterwards his time was increasingly
taken up with economic work.
In 1854 a vigorous attempt was made in Council to abolish
the gardens. They were defended by Mr. W. Ferguson in
the Observer ” and by Dr. Bindley in the “ Gardeners'
Chronicled Thwaites maintained the gardens in a very
high state of efficiency and opened out further portions of
the jungle for cultivation. The title of the post was changed
from Superintendent to Director in 1857, and in 1860 a new
European post was created, that of Conductor (now Curator),
to which Mr. Cameron of Kew was appointed. This
8
WILLIS : HISTORY OF THE
appointment relieved the Director of most of the routine
garden work, leaving him free to devote more time to the
work on the Ceylon Flora and to the introduction of new and
useful plants. In 1857 Thwaites obtained the assistance of
Dr. (now Sir) J. D. Hooker in the production of his proposed
Flora of the Island, and in the following y ear the first part of
it appeared, under the title of Enumeratio Plantarum Zey-
laniæ,” The work was completed in 1864, and contains the
descriptions of many new species, as well as notes on habitat,
uses, native names, &c. Thwaites at the same time issued
large sets of dried specimens of Ceylon plants (“ C. P. ” sets),
and with the proceeds of the sale of some of these added
many important works to the garden library.
In 1860 the site of the Hakgala Garden was selected by
Thwaites for the reception of the cinchona plants brought
from Peru by Sir Clements Markham, and was placed in
charge of Mr. MacNicoll. The garden lies on the side of Hak-
gala mountain at an elevation of about 5,600 fieet, in a cool,
temperate climate, with a mean annual temperature of 61°,
and rainfall of 91 inches, distributed over the year in
a similar way to that at Peradeniya. An area of about 550
acres of land was reserved for the garden ; most of this is still
covered with jungle or grass (patana). Here the cinchonas
were soon established, and in a few years large numbers of
plants were ready for distribution. At first planters were
disinclined to try the new industry, but a few years later,
as coffee ceased to do well, cinchona was very largely taken
up, finally becoming for some years the staple industry of
the Colony, and bridging over the period between coffee and
tea. The gardens began to distribute seed of the latter about
1864, but for many years previously Thwaites had called
attention to it in vain. In 1868 the new building of the
herbarium was opened, and the library and herbarium
moved into it from the Director’s house (the present Museum).
In the report for 1871 appears the first mention of the
coffee leaf disease, Hemileia vastatrix, and in all succeeding
CEYLON BOTANIC CARDENS.
9
reports Thwaites deals with this, consistently rejecting
the popular idea of the possible discovery of a cure. At this
period he devoted a good deal of attention to the fungi of
Ceylon, of which no less than 1,200 species, sent by him,
were described by Berkeley and Broome in the Journal of
the Linnean Society for LS71.
In the report for 1873 the spread of tea cultivation is
noticed, and the cultivation of cacao and cardamoms, both
now important industries in Ceylon, is for the first time
pressed upon public attention.
The “ Enumeratio ” was purely a scientific botanist’s Flora,
and with a view of getting more freedom from routine work
to enable him to compile a popular Flora of Ceylon, Thwaite®
obtained in 1874 the appointment of an Assistant Director
(the Conductorship had been abolished in 1863). The first
holder of this post was Mr. M. M. Hartog, now of Queen’s
College, Cork ; the second, Dr. D. Morris, now Imperial Com-
missioner of Agriculture in the West Indies. In spite of
this help, however, the duties of the Directorship proved too
heavy for Thwaites to carry out the proposed work, and it
was left in a very fragmentary condition at his retirement.
In 1876, the Indian Government having obtained seeds and
plants of Hevea hrasiliensis, the Para indiarubber, from
South America, it was found that the climate of India was
unsuitable for them, and they were sent to Ceylon, where
the branch garden at Henaratgoda, 17 miles from Colombo,
on the Kandy railway, was opened for their reception, and
as an experimental garden for strictly low-country products.
In this garden, with its hot steamy climate (mean tempera-
ture about 82°, elevation about 15 feet above sea level) and
well-distributed rainfall of 100 inches, many plants of the hot
equatorial regions flourish much better than at Peradeniya,
and the garden, besides its use as a nursery for rubber plants,
has proved a very useful adjunct to the principal garden.
It contains among other features of interest a small piece of
untouched jungle. The cultivated area is now about 30 acres.
49-01 (2)
10
WILLIS : HISTORY OF THE
Thwaites retired in the beginning of 1880, after a long
service of 31 years, spent without once quitting the Colony.
Dr. Morris, the Assistant Director, had left for an appoint-
ment in Jamaica a short time before, and the vacancy in the
Directorship was filled by the appointment of Henry Trimen,
M.B., F.L.S., Assistant in the Botanical Department of the
British Museum. Thwaites retired to Kandy, where he pur-
chased Fairieland bungalow. During his tenure of office he
had received the distinctions of F.R.S. (1864), Honorary
Degree of Ph.D., and C.M.G. (1878). He died in 1882 at
the age of 70. A memorial building, containing a brass, was
erected in the gardens soon afterwards.
The new Director arrived at a time of trial for the Colony,
for the coffee industry was rapidly sinking under the attacks
of its fungus enemy. At this period Mr. (now Professor) H.
Marshall Ward came out for two years, 1880-82, to carry out
a series of researches into the life-history of the Hemileia
with the view of endeavouring to discover preventive mea-
sures. The mission, though successful in the former object,
was unsuccessful in the latter, as Thwaites and Mr. Ward
himself had predicted would likely be the case. No Assist-
ant Director was appointed at this time, but the appointment
of Conductor was revived, and Mr. P. D. G. Clark was
appointed to it as Head Gardener. The garden at Peradeniya
was largely cleared of the redundant vegetation it contained,
and the South Garden, the last remaining uncultivated piece,
laid out systematically. With the appointment of Mr. W.
Nock in 1882 to the post of Superintendent of Hakgala, that
garden was also taken in hand, and gradually transformed
from a cinchona nursery into the beautiful and useful
Botanic Garden that it now is. In 1883 the small branch
garden at Anuradhapura, the capital of the North-Central
Province, was opened ; this gave the opportunity, hitherto
wanting, of growing many plants which are intolerant of
the wet climate of the three already existing gardens. The
climate of Anuradhapura, 300 feet above sea level, is that
CEYLON BOTANIC CARDENS.
11
of Peninsular India ; the mean annual temperature is over
80°, and the rainfall is chiefly in the last three months of the
year. Out of a total mean rainfall of 54 inches, 29 fall in
these months. In 1886 another branch garden of 11 acres
was opened at Badulla, the capital of the Province of Uva, at
an elevation of 2,220 feet. Here, on the eastern side of the
main mountain mass of the Island, the climate has a differ-
ent periodicity, and is also somewhat drier, than on the
western side. The mean rainfall is 79 inches, chiefly falling
in the north-east monsoon, and the “ dry ” season is from
July to September, instead of from January to March. The
flowering and fruiting seasons of many of the plants differ
correspondingly. The mean temperature of the locality is 73°.
In 1883 the Director’s bungalow in the centre of the
garden was transformed into a Museum, and gradually
stocked with a good collection of vegetable products and
specimens, chiefly economic. The Director’s residence was
transferred to the “ Assistant Director’s house, ” just vacated
by Mr. Ward.
Coffee was now failing fast, and Trimen devoted much
energy and enthusiasm to the introduction and spread of new
industries, especially cinchona, now becoming the staple of
the Colony, tea, cacao, and indiarubber, together with Liberian
coffee, and many others. He also devoted himself to the
preparation of a complete and thorough Flora of Ceylon, the
publication of which was at length begun in 1893. Unfortu-
nately he did not live to complete this work, but the last two
volumes of the five have been finished by the veteran Sir
J. D. Hooker, the concluding volume appearing in 1900.
Trimen also published, besides many purely scientific
papers, a Catalogue of the Plants growing in the Gardens, and
a Hand-Guide to Peradeniya Garden, of which four editions
appeared before his retirement. A fifth, re-written, appeared
in 1898.
About 1886 the garden began to attract visiting botanists in
appreciable numbers, and in 1888 a small room in the Museum,
12
WILLIS : HISTORY OF THE
next to the Director’s office, was turned into a laboratory for
their use, by aid of a grant made from the funds of the
British Association for the Advancement of Science. Mr. M.
C. Potter, now of the College of Science, Newcastle, was the
first to work in it. He was followed, in 1891, by Mr. J. B.
Farmer, now Professor at the Eo^^al College of Science,
London, and in 1893 by Mr. F. W. Keeble of Cambridge,
During the last few years of Dr. Trimen’s Directorship he
was mainly occupied with the preparation and publication of
his “Flora of Ceylon,” which forms a splendid monument of
his work here. His health failed very much in later years,
and on June 30, 1896, he retired on pension, remaining in the
Colony to work at the Flora, which unfortunately he was pre-
vented from completing by his death on October 16 of the
same year, in the 53rd year of his age. Hehadreceivedthe dis-
tinction of F.R.S. in 1888, and a brass to his memory has been
erected in the Museum by the Planters’ Association of Ceylon.
The writer, then Assistant in the University of Glasgow,
was appointed to succeed to the Directorship, and commenced
work on September 14, 1896. Mr. Nock had taken charge as
Acting Director since Dr. Trimen’s retirement.
During the last four years, which have been years of great
prosperity in Ceylon, owing to the growth and success of the
tea industry, many changes have been made in the Depart-
ment. Under the Curatorship of Mr. H. F. Macmillan,
appointed in 1895, the Peradeniya Garden has been greatly
improved, and a number of new experimental cultivations
started. There is not, however, enough space for the
efficient carrying out of real experimental cultivation of
important plants, and the Government has now under consi-
deration the opening of an Experimental Garden near to the
Botanic Gardens at Peradeniya, in which such work can be
properly carried on under a skilled Superintendent. With
this garden will be combined the training of students,
hitherto carried on at the Agricultural School in Colombo,
which was closed in March of the present year.
CEYLON BOTANIC GARDENS.
13
The tea industry, as well as others, having been troubled
with attacks of disease due to insects or fungi, the need of
having scientific help in the prevention and cure of such
troubles has been realized, and met by the appointment of an
Entomologist in 1899, Mr. E. E. Green, and of a Mycologist
in 1900, Mr. J. B. Carruthers. The latter officer also acts as
Assistant Director, taking charge of the whole Department in
the absence of the Director, and taking charge at all times of
certain branches of the economic work. A small annual vote
also provides for an Assistant (appointed for a period not
exceeding three years) for special investigations. The first
worker under this scheme was Mr. J. Parkin, who made a
very thorough investigation into the treatment of latex for
the preparation of indiarubber ; the present Assistant, Mr. H.
Wright, is occupied among other things with various ques-
ti ons bearing on the Forestry of the i sland, such as the sources
of the ebony and other Diospyros timbers, and the formation
»
of rings of growth in tropical timbers. Another important
appointment made in 1900 is that of Mr. M. K. Bamber as
Agricultural Chemist to the Department. Mr. Bamber has
so far been chiefly engaged in work on citronella oil, tobacco,
and camphor, the first two with the view of aiding established
industries, the last with that of aiding the establishment of
a new one.
The appointment of so many scientific officers of course
involved the construction of a proper laboratory, the more so
as the stream of scientific visitors from abroad was now
reaching considerable proportions. In 1896 and 1897 there
was only one visitor, but in 1898 the number rose to five, and
in 1899 to six. A new building, lying to the north of the
herbariupi, was commenced in 1899, and is now complete.
It contains rooms for general morphological, physiological,
chemical, economic, and photographic work, and has accom-
modation for eleven or twelve workers. Visitors from abroad
are cordially welcomed, and every facility provided for their
work. A resthouse is now being erected opposite the
14
WILLIS: HISTORY OF THE
entrance of the gardens, and there is ample hotel accommo-
dation in Kandy. In connection with the chief laboratory
at Peradeniya a small working room has been opened in the
garden at Henaratgoda, near which there is also a resthouse,
and a similar room, with sleeping and living accommodation,
is nearly completed at Hakgala. The Library of the Depart-
ment has been much enlarged, and now amounts to about
2,500 books and papers ; the chief periodicals are regularly
received.
The economic history of Ceylon during the last few years
is chiefly that of tea, which is by far the chief industry of the
Colony, though it deserves special remark that tea is not, in
proportion to other exports, nearly so great a staple as coffee
was. It forms about 52 per cent, of the value of the exports,
cocoanuts, cacao, and other products forming no small pro-
portion of the total. Mindful of the fate of coffee, the
Department makes every effort to encourage the development
of the minor industries, to introduce or discover new ones,
and to aid the planters in combating the first appearances of
disease among their crops. At present there seems every
prospect of indiarubber and camphor becoming valuable new
minor indu sr, ries in Ceylon, through the efforts of the Depart-
ment, which in the case of rubber date from 1876.
The present organization and staff of the Department is as
follows : —
Director : J. C. Willis, M.A., F.L.S.
Chief Clerk : R. H. Pereira.
Scientific Department»
Botanist : The Director,
Assistant Director and Mycologist : J. B. Carruthers, F.L.S.
Entomologist : E. E. Green, F.E.S.
Agricultural Chemist : M. K. Bamber, F.C.S.
Assistant : H. Wright, A.R.C.S.
Draughtsman : W. de Alwis, Muhandiram.
Herbarium and Laboratory Attendants, and two Plant
Collectors.
CEYLON BOTANIC GARDENS.
If)
Botanic Gardens Department.
Peradeniya. — Curator : H. F. Macmillan, and native staff.
Hakgala. — Superintendent : W. Nock, and nativestaff.
Henaratgoda. — Conductor : S. de Silva, and coolies.
Anuradhapura. — Conductor : D. F. de Silva, and coolies.
Badulla. — Conductor : D. D. Fernando, and coolies.
Proposed Experimental Gardens Department and
Agricultural School.
(Not yet organized.)
Superintendent, foremen, and coolies.
Teaching Staff : The Officers of the Department.
Note. — In the preparation of the above Paper I have made considerable
use of notes on the history of Ceylon Botany left by Dr. Trimen ; at some
later date I may hope to publish extracts from these in more detail.
Tlie Royal Botanic Gardens of Ceylon as
a Centre for Botanical Study
and Research.
BY
J. 0. WILLIS.
XT would be difficult to exaggerate the value of travel in
other countries to the working botanist, especially if
his work lie in the departments of systematic botany, geo-
graphical distribution, ecology, morphology, or economic
botany, whilst to the physiological or anatomical worker
there are also innumerable problems which can only be
solved by research in tropical countries. Even though only
a short visit be paid to a country whose climate and flora
are unfamiliar, and though no definite piece of research
work be undertaken, the traveller gains in breadth of view
and in understanding of the great problems of the science.
As the present reaction from the exclusive study of plants
in the laboratory to their study in the field as well gains in
strength, the necessity and advantages of travel and of study
and research in other countries will become more manifest.
The first flora and vegetation the traveller desires to see is
usually that of the tropical zone, and there is no place where
he can see it with greater ease than in Ceylon, nor any
tropical land with so great a variety of climate and vegetation
in so small a compass and so easily reached from whatever
headquarters may be selected. In the south-western plains,
between the mountains of the central part of the Island and
the sea at Galle and Colombo, the climate and the vegetation
[Annals of the Royal Botanic Gardens, Peradeniya, Voi, I., Pt. !.. June, 1901,]
49-01
(3)
18
WILLIS ; OBYLON AS A
are characteristically equatorial — a damp hot climate, very
equable, with rain at all times of year, and with the rank
luxuriance of vegetation that such conditions favour. Though
a large part of this region is now cultivated, there still remain
magnificent pieces of high forest ” into which botanical
excursions can be made. The north and east, on the other
hand, show a completely different climate and vegetation —
that of Peninsular India. The climate is dry for about nine
months of the year, and the forests show a totally different
composition and character. In the extreme south-east and
north-west, in the neighbourhoods of Hambantota and Man-
nar respectively, the drought is so extreme that the climate
and flora may be described as “ desert,” and very fairly com-
pared with those of parts of the Mediterranean region. In
the mountain ranges that occupy the centre of the Island
may be found every variety of climate and of flora from
tropical through sub-tropical and temperate almost to sub-
alpine, and both forest and grass land, both Avet districts and
dry, each with their characteristic peculiarities of flora. On
the southern and western shores of the Island may be seen
the characteristic Indo-Malayan shore flora, — the mangroves,
the beach-jungle (Barringtonia formation of Schimper), the
Ipomœa and the Nipa formations, &c. To the north the flora
of the coast changes to a more Indian type, and in the far
north may be found magnificent salt-lagoon floras. Water
plants abound in the coast lagoons, in the paddy fields, and
in the great irrigation lakes of the dry country, Podostemaceæ
in the mountain streams; sea-weeds are plentiful on the south
coast and in the great harbour of Trincomalee ; ferns, mosses,
liverworts, and fungi in the mountains and in the south-
western plains. Besides the rich flora of the Island itself
(over 3,300 species of flowering plants and ferns), the visitor
can in a very short time reach from Colombo any part of
India or the Malay Peninsula. The splendid jungles of the
Anamalais and other southern Indian mountains have been
but little explored, and can be reached in two or three days
from Colombo ; Assam, Burmah, and Malaya in a week.
CENTRE FOR BOTANICAL RESEARCH.
19
The traveller to Ceylon has the choice of numerous good
and swift steamship lines to Colombo. The four great
mail companies, the Peninsular and Oriental, the Orient,
Norddeutscher Lloyd, and Messageries Maritimes, run
steamers at frequent intervals from London, Bremen, Ant-
werp, Southampton, Marseilles, Genoa, and Naples. The
return fares are from London or Bremen about £80 first, £50
second class. The latter is quite comfortable, especially on
the French and German lines. A very popular line is the
Bibby,from Liverpool and Marseilles (return fare, first class
only, from Liverpool £74 10s., from Marseilles £69 10s.).
Other well-known lines are the British India, the City, the
Nippon Yusen Kaisha or Japanese line (noted for its cheap-
ness), and the Austrian Lloyd (from Trieste, return fare from
Trieste £42 10s., from London, including rail via St. Gothard,
£55). The fares to Colombo are high compared with those
to Australia, and the traveller with time at his disposal
should take a through ticket to Brisbane, costing a mere
trifie more, and break journey at Colombo, afterwards pro-
ceeding to the Straits, Java, or Australia.
It is a great mistake to suppose that any “ outfit ” is neces"
sary to come to Ceylon. All that is needed is a supply of
the ordinary clothing worn in summer in Europe. Khaki,
drill, and flannel clothes can be obtained, if needed, more
cheaply in Colombo than in Europe, and also sunhats and a
few necessary articles. There are good shops in the chief
towns of the Island, and almost everything may be purchased
as required at reasonable rates.
Living in Ceylon, though increasing in expense as else-
where, is still cheap enough. At hotels in the chief towns
it costs on the average about 7 to 8 rupees per day (the rupee
equals Is. 4,d. English money), but at resthouses* it is con-
siderably less, especially if the traveller take his own linen
* These are small furnished bungalows, with servants in charge, provided
by the Government in all important towns and villages and at intervals
of 14-15 miles along all chief roads, rendering travelling and botanising
in almost all parts of the Island a simple matter.
20
WILLIS: CEYLON AS A
(sheets, towels, &c.) with him. Hotel, resthouse, and bunga-
low servants speak English, so that there is rarely any need
for a personal servant, unless perhaps when travelling in out-
of-the-way districts. Travelling expenses are moderate ;
the most interesting regions in the Island, except the north,
may be reached by rail, and aline to the northern extremity
is now being made. Coaches run on the principal routes
not served by rail, and the roads are excellent for bicycling,
so that it is easy to travel by this method, sending heavy
luggage in advance by coach. Steamers run round the Island
weekly calling at Galle, Hambantota, Batticaloa, Trincomalee,
Point Pedro, Jaffna, Paumben, and Colombo.
The best time, in general, to leave Europe for Ceylon
is September, but, except for a somewhat hot three days in
the Red Sea, July or August is equally good. The stay in
Ceylon may last till March or April. April is a hot month
here, and after the middle of May the weather in the Indian
Ocean is usually very rough till the end of July. June and
July are wet but cool months at Peradeniya, and the weather
does not become at all unpleasantly warm till March or
April. The hotter weather may be easily avoided by going
“ up-country ” to work at Hakgala or elsewhere in the
mountains. There is, however, no reason for fearing the
heat at Peradeniya ; the temperature in the laboratory never
exceeds 84° F., and is usually between 75° and 80° (24-27° C)in
the middle of the day. Working hours are usually arranged
to suit the climate. A light “ early tea” at 6 or 6.30 A.M. is
followed by a walk in the garden, or collecting, and work in
the herbarium or laboratory from 7.30 or 8 till 11, when
breakfast is taken. Work may be resumed at 1 and continued
till 5 or after. There is no afternoon sleep in Ceylon as in
the Dutch colonies. Exercise (tennis, cycling, &c.) is usually
taken early in the morning or after 4.30 Dinner is usually
about 7.30. There is a good English Club in Kandy, and
also tennis and sports clubs, where cricket, football, golf, &c.,
may be obtained.
CENTRE FOR BOTANICAL RESEARCH.
21
Assuming that the absence from Europe is six months,
the cost of the trip need not much exceed £130, made up
thus : —
For first class on voyage add £20-35, and for more travel-
ling in Ceylon increase proportionately. The cost will be
proportionately less if a longer time be spent in Ceylon.
The best headquarters for the scientific visitor will in
general be at Peradeniya. A resthouse will shortly be
opened there, and other accommodation is often available.
The centre of Kandy town lies only four miles off, and there
are frequent trains from a station near the gardens. In Kandy
there are three excellent hotels, at which board and lodging
for visitors working at the gardens may be had at Rs. 5-6 per
day in periods of not less than a month at a time. These
rates do not include midday breakfast or tiffin, which is
assumed to be taken at Peradeniya. The resthouse at
Peradeniya lies immediately outside the garden gate, and
rooms in it will be specially reserved for scientific visitors.
The Botanic Garden at Peradeniya comprises an area of
over 140 acres (58 hectares), and contains a large collection
of tropical plants which are mostly carefully labelled.
Being laid out with large lawns and open spaces, the trees
are seen to advantage. In the centre of the garden lie the
buildings of the Scientific Department, comprising : —
(1) The Economic Museum, containing a very complete
collection of the Economic Products of Ceylon, systemati-
cally arranged ; in this building are also the offices of the
Director.
(2) The Library and Herbarium, containing on the ground
floor the offices of the Entomologist and Mycologist, the
Voyage, second class, and expenses on ship
Sixteen weeks at Peradeniya or other garden
Three weeks’ travelling
Miscellaneous necessary articles
£60
£ 130
22
WILLIS : CEYLON AS A
Lithographic Press room, and the duplicate collections from
the herbarium, and on the upper floor the Library, consisting
of about 3,000 books and papers and supplied with most of
the current botanical and economic literature of the day, a
complete and separate Ceylon herbarium, a general her-
barium, and a herbarium of the plants cultivated in the
gardens. The identification of plants in which visitors are
interested is thus rendered simple, especially as Ceylon
possesses a complete Flora, written by the late Director and
Sir J. D. Hooker.
(3) The Botanical Museum, not yet completed, but to
contain a Ceylon Botanical Collection, systematically
arranged,
(4) The Research Laboratory. This contains rooms for
general microscopic and morphological work (places for four
workers), for physiological work (two places), chemical work
(two places), economic (two places), and a private laboratory
for the Director, as well as a good photographic darkroom and
verandahs for experiments in the open air. It contains a
good supply of all the essential apparatus and reagents, but
workers from abroad must bring their own microscopes, or at
least lenses, and should bring what alcohol and bottles they
require for carrying away material. Such alcohol will be
admitted free of duty on certificate from the Director ; the
amount of duty must be deposited on entry, but is returned
when the alcohol is removed from Ceylon. To the labora-
tory is attached, besides the attendant, a plant collector?
whose services are within certain limits at the disposal of
visitors on payment of the cost. Coolies for special work can
be obtained on payment of their wages, usually about 37
cents a day, or Ç)d. in English money. Workers in physiology
or other lines requiring special apparatus should bring such
things with them, but for most work the laboratory is well
furnished with essentials.
The neighbourhood of Peradeniya is mountainous, and
mostly cultivated, but within a few miles there are some
CENTRE FOR BOTANICAL RESEARCH,
23
interesting pieces of wild country, and a short excursion by
road or rail brings many good collecting grounds within
reach.
Next to Peradeniya, the most important garden is at
Hakgala, about six miles from the great mountain sanitarium
of Ceylon, Nuwara Eliya. The garden lies on the steep side
of Hakgala mountain, at an elevation of 5,600 feet (1,680 m.),
and in a very good centre for exploring the mountain vege-
tation of Ceylon, On the west the hill jungles stretch away
in an almost unbroken sweep for 25 miles, and on the east
lies a vast expanse of dry grass country, with a peculiar and
interesting vegetation. The garden itself contains a large
reserved area of both jungle and grass (patana), and a very
interesting collection of plants, from Europe, Australia,
South Africa, the Himalayas, and other tropical mountains.
It contains a small laboratory with living accommodation
attached to it. The laboratory has two working places, and
contains a small local herbarium of the plants of the hills
and those cultivated in the garden. It has also a small
dining room and two small bedrooms, so that, by taking a
servant and arranging for supplies from Nuwara Eliya
twice a week, it is possible to live and work there in comfort.
A charge of one rupee a day is made for the use of the
building by visitors, who must supply their own linen and
knives and forks, but everything else is provided.
Another interesting station is the branch garden at Hena-
ratgoda, 17 miles by rail from Colombo on the Kandy line,
and reached in three hours from Peradeniya. There is a rest-
house close to the station, and the garden lies about | mile
away. It contains among other things a piece of untouched
jungle. The climate here is very hot and steamy, the
garden lying almost at sea level. Many interesting plants
are cultivated in it, and it contains a little working room,
with a few simple necessaries for botanical work ; anything
required can however easily be taken from Peradeniya.
From here an excursion may be made to the lagoon at
24
WILLIS: CEYLON AS A
Negombo to see mangrove vegetation, but this vegetation is
still more easily seen by taking the railway along the south-
west coast to Bentota (resthouse) or Ambalangoda (resthouse).
Nipa vegetation and many fine pieces of high forest may be
seen near Matara, the terminus of the line on the south
coast (resthouse), and the desert region of Hambantota
(resthouse) may be reached from here by coach.
There is another branch garden at Badulla (resthouse),
reached by rail and coach, on the eastern side of the
mountains ; the vegetation here is very interesting, and
excursions may be made into the low-country east of the
mountains, where the dry zone flora may be seen very well.
Another branch lies at Anuradhapura, in the north of the
Island, in the hot dry zone (resthouse), and may be reached
by rail and coach from Peradeniya. The town itself is
extremely interesting, being one of the famous “Buried
Cities of Ceylon.” In the neighbourhood there is interesting
jungle, and the flora of the great irrigation lakes is also of
interest. Another coach journey, soon to be replaced by
railway, to the north, brings the traveller to Elephant Pass
(resthouse), where the salt marsh flora may be seen to per-
fection, and beyond this lies the interesting Tamil country
of Jaffna, richly cultivated. From Jaffna the return may
be made by steamer, direct to Colombo, or to Trincomalee,
returning thence to Kandy by coach and rail.
While the laboratories and other facilities are primarily
intended for botanical work, there is no intention of exclud-
ing workers in other branches of science. Zoologists,
Entomologists, Geologists, and others will be cordially
welcomed, and places in the laboratories put at their dis-
posal so long as there is room. Of the eleven working places
in the principal laboratory at Peradeniya, about five are used
by the staff of the Department, leaving six available for
workers from abroad. Intending visitors should communicate
with the Director some months in advance when possible,
mentioning the line of work they wish to take up, and
CENTRE FOR BOTANICAL RESEARCH. 25
whether they wish any arrangements made before their
arrival, such as the planting of particulai* plants or seeds,
&c. The intending visitor would do well to read up before-
hand as much as possible about the flora and vegetation,
which will render it easier to carry on work here on arrival.
The following works may be specially mentioned : —
Tennent, J. E. : Ceylon. — Old, but very good.
Ferguson, J. : Ceylon in 1897,— Chiefly economic.
Cave, H. W.: Golden Tips, London, 1900.
Paris Exhibition Handbook to Ceylon, 1900.
Trimen, H., and Hooker, J. D. : The Flora of Ceylon,
1893-1900 (see especially the Introduction and
Supplements.)
Trimen, H. : The Flora of Ceylon, especially as affected
by Climate, Journ. Bot. 1886.
Trimen, H. : Remarks on the Composition, Geographi-
cal Affinities, and Origin of the Ceylon Flora,
Journ. Ceylon Branch, R. A. Soc., 1885.
Pearson, H. H. W. : The Botany of the Ceylon Patanas.
Journ. Linn. Soc. 1899.
49-01
(4)
CiilYL.O^
PRINTED AT THE GOVERNMENT PRESfc
COLOMBO
Fungus Cultures in the Tropics*
(^Preliminary Notei)
BY
CARL HOLTERMANN.
(With Plate I.)
N the introduction to my Mycologische Untersuchungen
aus den Tropen ” 1 have briefly indicated that during
my first visit to Ceylon I was successful in discovering a
nutrient substratum, which proved to be extremely favour-
able for cultures of fungi. I am anxious, before proceeding
to the publication of my recent work on the transpiration
relations of tropical plants, to give an account of this, in the
hope of thereby inciting others to test and further extend
my mycological results.
The nutritive solution which afterwards proved so useful
to me was accidentally discovered. At the commencement
of my tropical studies in 1895-96 I directed my attention^
in the first place, to the so-called Hemiascus fungi. Since
these were, according to my supposition, to be especially
looked for in the gummy excretions of trees, I allowed no
opportunity to pass of making investigations in this direction.
Excretion of gum from trees is a far more common pheno-
menon in the tropics than with us. In it there is always a
rich growth of bacteria and the lower fungi. In this
respect the well-known sugar palms, Arenga saccharifera
* “ Pilzkulturen in den Tropen,” translated by J. C. Willis at the
author’s request.
[Annals of the Royal Botanic Gardens, Peradeiiiya, Vol. I., Pt. II., December, 1901.]
28
HOLTE HMANN : FUNGUS CULTURES
and Caryota urens, are particularly remarkable. A reddish
brown gelatinous mass of varying size is, as a rule, to be
seen on wounded stems. This is the sweet sap, which
often flows in great quantity, especially from the cut ends
of young inflorescences. In it I always found the spores of
fungi, in one case of no fewer than twenty-six species, and
all were more or less germinated, some even with a large
mycelium. The species found were Basidiomycetes,
Mucorineæ, üstilagineæ, and others. My first attempts to
use this exudation for fungus cultures were without success,
and need not be described here.
The sap of Arenga saccharifera and Caryota urens is dried
by the natives and brought to market, without further
purification, as palm sugar, consisting of sugar, vegetable
acids, and salts. It was a dilute solution of this in water
which proved so excellent a medium for the cultivation of
different genera of fungi. Palm sugar is everywhere
common and cheap in Java and Ceylon, and is not difficult
to obtain in Europe. The various genera of fungi require
different degrees of concentration of the medium. Some
fungi germinate easily in a 25 per cent, solution, others
only in one of 5-6 per cent. Definite rules can hardly be
laid down. I have been in the habit, when the spores will
not germinate in a 10-11 per cent, solution, of trying a
weaker one.
The sap, being an excretion product of the plant, naturally
contains salts and other chemical combinations besides
sugar, and it is to this that its great advantages as a nutritive
medium are to be attributed — sugar alone would not be so
suitable. Only in rare cases must anything be added to it
to ensure germination.^'
On account of the impurity of the Java and Ceylon sugar,
the watery solution must be filtered before use, and various
measures must be taken to free it from the numerous
E.g., see my ‘‘ Mycologische üntersuchungen,’' p. 20.
IN THE TROPICS.
•29
foreign micro-organisms. Of the different well-known
sterilization methods, it is, as a rule, sufficient for mycologi-
cal purposes in the tropics to use the so-called fractional
method, which was invented by Tyndall — the boiling of
the nutritive medium four or five times, with an interval of
one or two days between each two operations. This, how-
ever, is not always sufficient for the above-mentioned palm
sugar solutions. A bacterium occurs regularly in the
solution of Arenga (not of Caryota)"^' sugar, which has a
great similarity to the common hay bacillus, and which is
distinguished by the great capacity of resistance of its
germs. It is particularly resistant to heat, and makes the
sterilization of the solutions — a necessary preparation for
the success of mycological cultures— very difficult. I
endeavoured repeatedly to confine myself to the convenient
fractional sterilization method. Although the boiling in
glass flasks was repeated for several weeks, and the tem-
perature raised to 100° C. on the first occasion, I was not
successful in attaining my object — the bacteria appeared
again and again, and as there were no other organisms to
affect their successful growth, they soon took possession of
the entire fluid and made it quite unsuited for cultures.
The other micro-organisms, though at first often present
in large quantity, were more sensitive to outside influences,
and a heating three or four times repeated in four or five
days was enough to kill them.
In the case of the nutritive solution under consideration
it is in general best to kill the resistant spores of the
bacteria by heating to about 110° R. in an autoclave. After
this process has been once employed, it will suffice to
employ a fractionated sterilization to 70° R. for the three or
four further heatings. The solution can indeed be easily
freed from all germs by an immediate heating and the first
necessity for a pure culture be thus obtained, but experience
shows that its nutritive capacity is thus seriously lessened.
For clearness’ sake I repeat of Arenga (not of Caryota).
30
HOLTBRMANN : FUNGUS CULTURES
As has already been mentioned, the solution contains,
besides sugar, other organic compounds, including proteid.
The proteid is not indeed present in such quantity as
to coagulate on heating, but its composition is apparently
altered, for in the solution sterilized only by means of the
autoclave many spores failed to germinate.
Just as is the case with the bacteria, there is no universal
nutritive solution for fungi. It has long been well known
that many spores, e.^., those of Coprinus and other copro-
philous fungi, can only be cultivated on a decoction of
dung. It has been noticed, in the course of my experiments,
that a considerable number of dung-inhabiting fungi will
germinate easily in my solution, if there be added to it a
small proportion of meat extract, about 2 gm. to 50 gm.
of the solution. The mycologist in the tropics cannot be
too strongly advised to prepare a dung solution in Europe
and take it with him for all purposes. I have easily
succeeded in making such a solution in the Berlin Institute
under great pressure (3-4 atmospheres) in an autoclave.
The method has been long since described by other authors ;
horse dung is placed in a vessel with about one-third of its
weight of water, and heated for about a quarter of an hour
under the pressure mentioned. The solution is filtered and
sterilized in the flasks to be described below. In these it
can be easily carried without fear of contamination.
Provided with these two solutions it is possible to under-
take in the tropics many interesting developmental investi-
gations of the fungi.
There are many fungus forms in the tropics which cannot
be brought to germinate by the means at present available,
e.^., the Phalloideæ, most Clavarieæ, the red-spored Basi-
diomycetes, &c. With these I had no success. A satis-
factory explanation of this passive behaviour towards all
the nutritive solutions employed is not at present forth-
coming. Nevertheless, I would urge further investigation
IN THE TROPICS.
31
into the germinative capacities of fungus spores, and
do not doubt that a solution may be found suitable for
the growth of these forms, which have as yet proved
resistant to all methods tried. The period of germination
varies much ; some spores, such as those of the Muco-
rineæ, form long hyphæ after a few hours, while others
only show the first phenomena of germination after three
or four days. According to my experience, spores which
remain unaltered after a week in the solution usually
remain permanently in this condition. A certain amount
of foresight must also be exercised in the choice of spore
material. A few fungi, especially among the Myxomycètes,
seem to lose their capacity for germination in a few days,
and this again may vary in the same species. For instance,
I found an Ustilago very common in Java on a Polygonum :
the fresh material only germinated very slowly, while later
experiments tried in Berlin succeeded admirably. In
Arcyrea punicea (a Myxomycete), on the other hand, the
fresh material was preferable, for in the course of only an
hour and a half after sowing the slide was full of swarm
spores, while a few months later the results were entirely
negative, and no efforts would induce germination of the
spores, which formerly had proved so easily manipulated.
Other species, again, retained their capacity for germination
for a long period, but became sluggish, ^.e., their germination
was later. In many i^scomycetes in particular I frequently
found that freshly collected material was most satisfactory,
but that after the lapse of 6-12 months the germination was
much slower. In the Myxomycètes the spores particularly
soon lose their capacity for germination. As is well
known, the growth of the higher plants in the tropics is
much more rapid than with us. Judging from cultivation
experiments this is not the case with the fungi, or at any rate
the germination and the subsequent formation of mycelium
proceeded, with the same species, just as rapidly in the
Berlin Institute as in Colombo or in Batavia. It was also
32
HOLTERMANN : FUNGUS CULTURES
impossible to find any distinction in this respect between
Colombo and Nuwara Eliya, although the mean temperature
in the former was 27° C., in the latter only 15° C., and the
experiments were carried on in both places with the same
species (Schizostega, Dacryomycetes, and Mucorineæ).
As has just been pointed out, many fungi can never be
germinated in the usual nutrient solutions. To my great
disappointment this was also partially the case with Hemi-
leia vastatrix, a fungus which has, as is well known, caused
immeasurable damage to coffee trees ; it has almost totally
destroyed the cultivation of coffee in Ceylon and other
colonies by injuring the leaves. I take this opportunity of
describing my cultivation experiments in the hope that
others may carry them further. The yellowish spores occur
in great numbers on the under sides of the leaves. Their
detailed description is already sufficiently well known.
The outer coat of the spore is covered with small papillæ,
between which, as a rule, small particles of dust adhere to
the spore, so that it is in consequence a matter of great
difficulty to obtain pure spore material ; but success may be
attained by causing the spores to fall into a drop of a
mixture of water and alcohol by giving a gentle fillip to the
leaf. The solution should contain only about 6-8 per cent,
of alcohol, and must be made with sterilized water. It
frequently happened that even in this very weak solution of
alcohol young spores completely lost their capacity for ger-
mination. It is necessary to work with great care to prevent
the influence of the solution from being too intense.
The first stages of germination showed themselves in my
experiments after a couple of hours, one or two hyphæ
appearing. These grew longer and longer, and finally grew
upwards out of the solution into the air. Great masses of
hyphæ appeared there, proportional to the number of spores
sown. In other cases these remained upon the solution,
and there showed much branching. It is of special interest
to note that the hyphce never showed any transverse walls.
IN THE TROPICS.
33
As is well known, it has been shown by De Bary that this is
only the case with the lower fungi. Hitherto Hemileia has
been counted among the Uredineæ, but whether this is
really the case must remain very doubtful after the obser-
vations above mentioned, for, as is well known, the hyphæ
of the other Uredineæ possess distinct transverse walls. It
is further to be noted that the hyphæ always remain sterile ;
in spite of all attempts I failed to obtain any form of fructi-
fication ; as long as the culture remained alive, it continued
sterile, although at times small swellings appeared at the
tips ; it is probable that these would have developed into
spores had it been possible to keep the cultures alive longer,
but as they were always killed through contamination with
Bacteria and Pénicillium, I was at the most only able to
study them for a fortnight. The spores appear only to
germinate in presence of plenty of free oxygen, for the
above-mentioned phenomena of germination only appeared
in those which were swimming on the surface of the solution,
those which were submerged all remaining unaltered.
Although these cultivation experiments afford but little of
interest, I have taken the opportunity of describing them,
in order to point out that it is certain that with modern
methods important contributions may be made to the deve-
lopmental history of Hemileia ; the principal difficulty is
to obtain pure cultures, and for this there are still many un-
tried methods available ; once let the developmental history
pf this fungus be made out, and we can then hope to find
means of limiting the devastations of this formidable disease.
The nutritive solution was always used in a very dilute
form ; the best results were obtained with solutions which
had not been heated for some months.
Sterilization naturally plays a great part in mycological
investigations in the tropics. In the rainy seasons almost
everything is infested with fungi ; I have always and in all
cases used alcohol. Brightly polished instruments were
kept in spirit, and before use passed through a flame, and
34
HOLTERMANN : FUNGUS CULTURES
coverslips were kept germ-free in the same way. Micro-
scope slide glasses were taken out of the alcohol with forceps
and at once passed through a flame ; if they are then laid
upon one another in a sterilized vessel they can be pre-
served for a long time. Flasks, after careful washing,
were thoroughly washed out with alcohol and stood bottom
upward, but in such a way that the alcohol could evaporate,
and afterwards closed with a plug of wadding, which had
been previously freed from germs with a solution of corro-
sive sublimate in alcohol (1 part in 1,000). Other means of
disinfection for glasses and instruments are not to be recom-
mended ; I must particularly utter a warning against the
use of carbolic acid, corrosive sublimate, and hydrochloric
acid, as this involves the risk of spoiling the culture experi-
ments. In Europe, where as a rule all kinds of drying
ovens, autoclaves, &c., are available, other methods may
indeed be used, but even here alcohol is to be recommended
to the mycologist in most cases. It is obvious that by this
method all such articles as corks, filter papers, &c., which
cannot be exposed to high temperatures, may easily be freed
from germs. The remarkable efficacy of alcohol has long
been known. Some authors indeed consider even a 14 per
cent, solution as sufficient ; in the tropics this is, however,
often not strong enough to kill the bacteria. The bacterium
above mentioned as occurring in palm sugar showed itself
particularly resistant, and only lost its germinative capacity
when the solution reached a strength of 20 per cent. In
general, 1 have used 75 per cent, alcohol, in which the germs
are at once destroyed.
When we have thus freed the nutritive solution from all
germs, and so fulfilled the first condition for the establish-
ment of a pure culture, every possible care must be taken to
preserve it from any subsequent contamination from with-
out. For this purpose I recommend a specially constructed
flask, which is illustrated in the figures. It is an ordinary
washbottle, whose blow-tube (a) is blown out into a
IN THE TROPICS.
35
spherical bulb, which is filled with wadding. The other tube
(b) is drawn out to a point and covered by a protecting
tube (c). The jet must be so fine, that when used it shall
only give one or two drops ; fungus cultures being usually
made on the slide in a drop of nutrient solution, the flask
must not furnish too great a quantity. In order to keep out
particles of dust the protecting tube (c) is closed with a plug
of wadding damped with alcohol. The wadding in the blow-
tube (a) is sterilized with corrosive sublimate before being
used. Wadding is an excellent bacterium filter in Europe
without further treatment, but in the tropics, where the air
is so damp, especially in the wet season, mould fungi are
easily able to get through it ; in a few days the ubiquitous
spores begin to germinate and send their hyphæ through
the thick tissue of the wadding, and in a very short time
the under side is quite green with the plentifully formed
conidia of Pénicillium. Against this danger the sublimate
is a protection. The cork or rubber stopper of the flask
must be protected in the same way against the entrance of
spores or hyphæ. If these precautions are carefully attended
to, the flask may be used for years without any fear of con-
tamination.
It is obvious that the nutritive solution should be steri-
lized only after it has been introduced into the flask, and
that the wadding in the b'ow-tube (a) should not be inserted
till after this has been done, otherwise the steam will carry
particles of the corrosive sublimate into the solution. The
cork of the flask (d) may be fastened down with a thread to
keep it permanently in one position, and the flask should
never be opened without real necessity, and for a short time
only. In this case care should be taken to hold the flask
erect, to hinder the entrance of the countless spores of the
tropical air. A fresh sterilization is to be recommended after-
even a single opening. It is also necessary to examine the
fluid microscopically from time to time to make sure that it
is completely sterile and contains at most only dead germs.
(6)
38 HOLTERMANN : FUNOUS CULTURES
Modern mycology operates in Europe principally with
very fluid nutrient solutions. As is mentioned on page 2 of
the introduction to my ‘‘ Mycologische Untersuchungen,”
such were but little used in my work. For as the cultures
even in moist chambers are easily exposed to evaporation,
the nutritive solution must be supplied in a form in which
its concentration and composition may remain unaltered. I
employed an addition of various Tremella or Auricula decoc-
tions with good results. The mass appeared almost as clear
as water, and possessed more or less viscosity according to the
concentration, but, finally, I was successful in the hottest
season atBuitenzorg with a little gelatine. The nutritive sub-
stratum is not rendered solid by the addition of the gelatine
as in our climate, but in the warm tropical regions remains
always more or less viscous. To prepare a firm substratum
it is necessary to add a mixture of agar-agar with gelatine.
Obviously, in employing viscous fluids, the jet of the flask (c)
must be correspondingly enlarged, or better, the flask may be
slightly warmed till the solution has the necessary fluidity.
The influence of daylight on the cultures, especially in the
tropics, is very injurious, and they must therefore be kept
in semi-darkness. The injurious effect is to be chiefly at-
tributed to the fact that the solution loses water in light,
thus producing an unsuitable degree of concentration. It is
the more necessary to protect the cultures from evaporation,
as a subsequent addition of solution is disadvantageous, the
currents thus induced often affecting the culture. To ob-
tain absolutely pure cultures is an operation attended by
great difficulties, and in the tropics particularly, if errors are
to be avoided, it is necessary to work with extreme care and
cleanliness. A culture should contain only one species. A
strange germ is capable within a short time of altering and
destroying this. Certain fungi, e.g.^ some'Dacryomycetes,
showed themselves, however, capable of overcoming all intru-
ders, even Pénicillium, which as a rule destroys everything.
Botanical Institute of the Royal University,
Berlin, Autumn, 1901.
Aim. Per ad. I.
PLI
C.Holtermam iel.
Lith.ÄrLSt:v,E.li’unie; leipzi(j.
TN THE TROPICS.
37
Explanation of Figures in Plate I.
Fig. a-d. Flask for nutritive solution, x
Figs. 1-7. Hemileia vastateix.
1. Germinating spores in nutritive solution, 1-24 hours after
sowing. X 230.
2. The same spores after 36 hours, x 230.
3. Germinating spores, 48 hours after sowing, x 230.
4. Mycelium, 14 days old. x 300.
.5. Mycelium from a coffee leaf, showing transverse walls ; the
only case that I have observed, x ISO.
6. Very young, unripe spores. Exhibited slight germination
phenomena in the nutritive solution, but soon died, x 230.
7. Spores on a coffee leaf, x 180.
V*, • N S
,w Ê
»‘^•w .i'
'.Av,' ■:' '
,V ’»k
« '-
{•■.H r
,r(
'1
W ,'
: «
■ ,.î-'
. ^
’
./. 't-
w" ' ;. ■“ >... r
■i
. M
!)
■;;?‘v.:h.:î i ■■ ^
'- ä£..
/ ■'’«!' nJ'”'
v^r: ■ '{'■ V--
>,- ■- ,' ;■' .’-i
., var. Acida, Brandis ; Lime, Limbo, M. ;
Limbu, Hind. Generally cultivated. — J. S. G.
Cultivated in the tropics. Ameni, Anderut, Kiltan (Lacca-
dives), Minikoi.
Var. LIMONÜM ; Lemon ; Bodu Limbo, M. (=big lime).
Cultivated in Landu.— J. S. G.
Aurantium, L, ; Orange ; Moli, M.
Cultivated in Fua Muluku, from which the fruit is some-
times brought to Male.” — J. S. G.
Cultivated in warm countries. Anderut (Laccadives).
Decumana, Z/. ; Niyaduru, M. ; Shaddock, Pumelo.
I. Didi, 25 ! Trimen’s list. Male and Addu, Gardiner.
Cultivated in Trop. Asia, &c. One tree on Minikoi, Prain.
[Ægle Marmelos, Gorr. ; the Bael. Cultivated in Mi-
nikoi, Prain.]
Simaruhece,
Suriana —
maritima, L.; Halaveli, Halia veli, M.
I. Didi, 137 ! Goifurfehendu Atoll, Hulule. At the ex-
treme S.W. corner, a little w^ay back from the beach ;
sparsely distributed in the group, never in large masses,
Gardiner !
A cosmopolitan littoral tropical species ; rare in Ceylon,
known only from Jaffna and Trincomalee ^ Minikoi, Prain.
Laccadives, in Bitrapar, Bangaro, and Kadamum, Prain.
Diego Garcia, Hemsley.
Meliacece.
Azadirachta, a. Juss,
INDICA, A, Juss. ; Hiti, M. ; Kohomba, S. ; the
Margosa.
I. Didi, 64 ! “ Not seen.’W. S. G.
Native of India and Ceylon, commonly cultivated for the
sake of the medicinal bark and oil (obtained from the seeds).
(9)
60 WILLIS AîîD GARDINER • BOTANY
[Carapa moluccensis, Lam.; occurring on shores of Ceylon,
Africa, Asia N. ; might occur in the Maldives.]
[Celastraceæ. Pleurostylia Wightii, PT. cfc A.; in Kada-
mum (Laccadives), Prain.]
Rhamnacece.
Zizyplms, L.
Jujul)a,Lam. ; Kunnaru, Konara, M. ; Maha-debara,
S. ; Ilantai, T. ; Kumara, Persian.
I. Didi, 147 ! Male, Christopher, 1888 ! Goifurfehendu
Atoll, Hulule 16, Veimandu in Kolumadulu Atoll, Gardiner !
‘‘ Cultivated and common in the whole group, a few trees
in every village. The fruit is eaten.” — J. S. G.
Native of, and largely cultivated in, India and Ceylon.
Probably introduced for cultivation, especially as it is absent
from the Laccadives. Minikoi ?
sp. ?
A fragment from Hulule, collected by F. Lewis, is a
second species of Zizyphus.
Colubrina, Rich,
asiatica, Brongn, ; Ra-rohi, M. ; Tel-hiriya, S.
I. Didi, 122 ! Turadu in S, Mahlos, Hulule 27, Gardiner !
Covers nearly the whole of the centre of the eastern side
of Hulule, but not on the beach ; not common in other
islands, except in Limbo Kandu (N. Mahlos).
A littoral species common in S. Africa, the Mascarene Is.,
S. India, N. Ceylon, Malaya, N. Australia. Not recorded
for Minikoi ; Laccadives, in Akati and Kiltan, Prain.
Ampelidece,
Titis, L,
Linnæi, Wall?
Malé, Trimen ; no specimens in the herbarium at Pera-
deniya. “ Not seen.”— J. S. Q. •
OF THE MALDIVE ISLANDS.
61
The occurrence of this species must be regarded as doubt-
ful. Perhaps bird introduced. Native of dry regions of
Ceylon and S. India.
[V. quadraiigularis, Wall^ in Kalpeni, T. carnosa, Wall, in
Kiltan (Laccadives), Prain.]
Sapindaceœ.
Cardiospermum, L,
Helicacabuni, L.
Small island off Malé, F. Lewis ! Cosmopolitan, tropical.
Laccadives, in Ameni and Kalpeni, Prain,
Allophyliis, L.
Cobbe, Bl.\ Dorn Moussa, M. (Dorn Musa in Minikoi) ;
Kobbe, S.
I. Didi,* 167 ! Hedufuri, Goifurfehendu Atoll, Hulule 26
and 40, Veimandu in Kolumadulu Atoll, Gardiner ! “ In the
densest jungle on Hedufuri.” — J. S. G. “ A big tree by the
village on Hulule, introduced by Dorn Moussa.” — J. S. G.
A littoral and inland species, common in India, S.W. Ceylon,
Malaya, N. Australia. It may quite possibly be native in
the Maldives, perhaps introduced by birds or sea, but was
most probably introduced artificially. Minikoi, Prain, Gardi-
ner ; probably, as the name there is the same, and quite
unlike the Sinhalese or Tamil names, this plant was intro-
duced there from the Maldives. It can hardly be accepted
as an undoubted native for either place, especially as it is
not found in the Laccadives. The fruit is edible.
Dodoiiaea, L.
viscosa, L,; Kudiruvali, M.; Eta-werella, S, ; Virali, T.
I. Didi, 139 ! Goifurfehendu Atoll, Veimandu in Kolu-
madulu Atoll, Gardiner ! Very characteristic of rocky
jungly area in most of the Archipelago.
A cosmopolitan tropical shrub, both in dry inland spots
and on the coast ; probably sea-borne, Minikoi, Gardiner,
35. Not recorded for the Laccadives,
62
WILLIS AND GARDINER. : BOTANY
ANACAEDIACEÆ.
Mangifera, L.
Indica, L. ; Amba, S. ; the Mango.
“No trees ; a few young plants in the Malé. The intro-
duction of this tree has been tried many times, but it will not
flourish.”— J. S. G.
Universally cultivated in the tropics. One tree in Mini-
koi, Prain (“None now.” — J. S. G.).
MOR1NGEÆ.
Moringa, Burm.
Pterygospbrma, Gaertn.; Muranga, M.; Murunga,
S. ; the Horse Radish Tree.
I. Didi, 4 ! Hulule, Veimandu in Kolumadulu, Gardiner !
Common in all the group near the villages ; seeds, &c., used
in curries. — J. S. G.
A native of N. India ; much cultivated in India and Ceylon.
Minikoi, cultivated, Gardiner, 10 ; Laccadives, cultivated
in Ameni and Kiltan, Prain.
Leguminosce.
Crotalaria, L.
retlisa, A.; Viha-giguni, M.; Kaha-andana-hiriya, S.;
Kilukiluppai, T.
I. Didi, 31 ! Kaddu in Haddumati Atoll, Gardiner !
A native of Tropical Asia and Australia, now a cosmo-
politan tropical weed. Laccadives, in Akati, Prain.
[C. verrucosa, L. ; in Ameni and Kadamum (Laccadi ves),
Prain.]
Indigofera, L.
tinctoria, L. ; Vihafilia, M. ; Nil-awari, S. ; the
Indigo plant.
1. Didi, 77 ! Hedufuri, Goifurfehendu, Veimandu in
Kolumadulu, Gardiner ; not cultivated ! Male, Trimen.
OP THE MALDIVE ISLANDS.
63
Universally cultivated in India, and probably originally
introduced with a view to cultivation."* Laccadives, in Kada-
mum and Akati, Prain.
[I, cordifolia, Heyne ; in Kiltan and Kadamum (Lacca-
dives), Prain.]
Tephrosia, Pers.
tenuis, Wall ; Fesko, M.
I. Didi, 10 (Prain) !
A cultivation weed, native of N.W. India. Laccadives, in
Kadamum and Akati, Prain.
purpurea, Pers., var. pumila. Baker.
Hedufuri, Gardiner ! Frequent in paths in low jungle ;
not in Addu.— J. S. G.
A cosmopolitan tropical weed. Laccadives, in Kadamum
and Kiltan, Prain.
[Sesbailia aculeata, Pers.; in Kalpeni (Laccadives), Prain.]
[Sbsbania grandiflora, Pers.; in Kiltan, Kadamum,
Akati (Laccadives), and in Minikoi, all cultivated, Prain.]
[Arachis HYPOGÆA, L.; the Ground Nut; cultivated in
Minikoi, Prain.]
Desmodium, Desv.
umbellatum, DC.
Goifurfehendu Atoll, Gardiner. On rocky land. Very
common in the Archipelago.— J. S. G.
Ceylon, Mascarene Is., Burmah, Malaya, Polynesia. Per-
haps sea-borne.
gangeticum, DC.
Hedufuri, Gardiner (Prain) !
A weed, Tropical Africa and Asia, Ceylon. Possibly sea-
borne.
triflorum, DG,; Hekoopie, M. (Minikoi) ; Hin-undu-
piyali, S.
* Probably so : it was certainly very common in Hedufuri and the
other, islands in S. Mahlos, where cloth is extensively made, and was
probably introduced for its dye = cloth or thread, while feli is a
particular g'arment) — .J. S. G.
WILLIS AND GARDINER: BOTANY
Btiruni in Kolumadulu, Gardiner ! Never seen elsewhere ;
grows on the broad mass of roots of the cocoanut tree.
Cosmopolitan tropical littoral and inland weed. Lacca-
dives, in Akati, Kiltan, and Kadamnm, Prain. Minikoi,
Prain, Gardiner. Perhaps sea-borne.
(Mucuna capitata, W. ê A.; in Ameni (Laccadives),
Prain.)
Erythrina, Linn.
indica, L.; Berebedi, M. ; Erabadu, S.
I. Didi, 13 (Prain) !
India, Ceylon, Burmah, Malaya, Polynesia. Probably
intentionally introduced, but perhaps sea-borne. •
Canavalia, DO.
Ensiformis, Z)(7., cultivated var, ; Talafuri, - M.
(Minikoi).
Veimandu in Kolumadulu, Gardiner (Prain) ! On waste
land, not on the beach.
Cultivated and more or less wild in India and Malaya.
Minikoi, Gardiner. Not recorded for the Laccadives.
lineata, DC.{oltusifolia^ DG., of Fl. Br. /.); Manifa, M.
Hedufuri, Hulule 13, Gardiner (Prain) !
Coasts of Tropical Asia, Minikoi, Prain. Sea-borne.
(turgida, Grah.; in Minikoi, Prain.)
Phaseolus, L.
Lunatus, A.; Himeri, M.
I. Didi, 54!
Cultivated all over the tropics. Not recorded for the
Laccadives or Minikoi.
(calcaratlis, Roxh.; an escape in Minikoi, Prain.)
Vigna, Ravi.
lutea, A. Gray.
Veimandu, Gardiner (Prain) !
A littoral species, cosmopolitan in the tropics ; not in
India or Ceylon, Minikoi, Prain,
OF THE MAEDIVE ISLANDS.
'65
[Catiang, EndU ; cultivated in Ameni and Kadamum
(Laccadives), Prain.]
Pachyrhizus, Rich.
AngulatuS, / Yiha-toli, M.
I. Didi, 96 !
Universally cultivated in the tropics.
Clitoria, L.
Ternatba, L.
Horsburgh Atoll, Gardiner! By mosques in Goidu, and
in the south. — J. S. G. Ceylon and the tropics generally,
often cultivated. Probably introduced intentionally.
Cultivated in Ameni, Laccadives, Prain.
Dolichos, L.
Lablab, L.; Himerri, M.
Hulule 39, Kaddu in Haddumati, Gardiner (Prain) I Male,
Trimen’s list ; no specimens. Cultivated all over the tropics.
Probably introduced for cultivation. Not recorded for the
Laccadives or Minikoi.
Sophora, L.
tomeiitosa, L.
Goifurfehendu Atoll, Gardiner ! Not seen elsewhere ex-
cept at Male. — J. S. G.
Cosmopolitan on tropical coasts ; sea-borne. Ceylon, India.
Not recorded for the Laccadives or Minikoi.
Cæsalpiiiia, L.
Bonducella, Fleming ; Karikuburu, M.
Hedufuri, Goifurfehendu, Veimandu in Kolumadulu,
Gardiner ! Didi, 60 ! Common on sandy shores behind
the Scævola in the whole Archipelago.
Cosmopolitan in the tropics, on the coast and inland.
Not recorded for Ceylon or Minikoi. Laccadives, in Bangaro,
Kadamum, Akati, Prain.
Bonduc, Roxh.f
A fragment found among the former species appears to
be this ?
66
WILLIS AND GARDINER : ROTANT
PULCHBRRIMA, ; Patangu, M.; Peacock Flower.
I. Didi, 51 !
Cultivated for its ornamental flowers in all tropical coun-
tries.
POINCIANA, L,
Regia, Boj, ; Kandi-toli, M. ; Flamboyante.
I. Didi, 3, a fragment, appears to be this. “ One tree in
one of the Sultan’s house compounds, and one by a mosque
bathing pool near the W. of Male.” — J. S. G.
Cultivated in all tropical countries for ornament.
Cassia, L.
occidentalis, L,; Kuhada, M. (Minikoi) ; Penitora, S.
Miladumadulu, Gardiner (Prain) ! Malé, Capt. Simons !
A cosmopolitan tropical weed. Minikoi, Gardiner ; Lacca-
dives, in Akati, Prain.
Sophera, L. ; Rana-rua, M.; Ura-tora, S.
Goifurfehendu, Hulule, Male, Veimandu in Kolumadulu,
Gardiner ! Malé, Trimen. No specimen in the Peradeniya
herbarium. “Grown as an ornamental plant round the
mosques, especially in Addu and Suvadiva ; in the latter it
is reported to be connected with the Shastras.” — J. S. G.
Cultivated and semi-wild.
A cosmopolitan tropical weed. Not recorded for Minikoi
or the Laccadives.
Tora, L,
Goifurfehendu, by mosques, Gardiner !
An almost cosmopolitan tropical weed. Laccadives, in
Kalpeni, Kadamum. Akati and Kiltan, Prain.
auriculata, L. ; Ranawia, M.
Malé, Trimen. No specimens.
glauca, Lam., var. sufTruticosa, Koenig ; Rana-ura, M.
I. Didi, 105 (Prain) !
A weed ; India, Ceylon at Trincomalee ? Malaya, Aus-
tralia, Polynesia.
OF THE MALDIVE ISLANDS.
67
Tamarindus, L.
Indica, L, ; Helebeli, M. ; Siyambala, S. ; Puli, T. ;
the Tamarind.
I. Didi, 44 ! Horsburgh Atoll, Gardiner ! Sparsely near
villages everywhere, common in Addu. — J. S. G. Trimen ;
no specimen.
Cultivated in the eastern tropics. Laccadives, in Akati
and Ameni, Prain.
(Afzelia biglija, A. Gray ; in Diego Garcia ; found in the
Seychelles, Madagascar, Malaya, Polynesia.)
Parkia, R. Br. ?
I Didi, 58, a fragment, with the native name Bes-gobili
{cf. Acacia), seems to be a Parkia, sp. ; planted in all pro-
bability as a shade tree, as in Ceylon.
Adenaiithera, Roy en,
pavonina, L.
Trimen’s list ; no specimen seen. Perhaps planted.
[Entada scandeiis, Benth,; seeds were found on Hulule
beach by Mr. F. Lewis.]
Mimosa, L.
pudica. L.; Ladu, M. ; Nidi-kumba, S. ; Lajalu, Hind.
I. Didi, 21 ! “Only seen near E. end of Male, where there
is a single clump.” — J. S. G.
A cosmopolitan weed in the tropics. Not recorded in
Minikoi or the Laccadives.
Acacia, L,
Fariiesiana, Willd. ; Bes-gobili, M. ; the Cassie Flower.
I. Didi, 127 ! Goifurfehendu, Veimandu in Kolumadulu,
Gardiner ! Common in the south of the group by mosques
and in the jungle, rare in the north.
Cosmopolitan in the tropics, often planted ; naturalized
in Ceylon. Not recorded for the Laccadives or Minikoi;
possibly intentionally introduced in the Maldives, but more
probably native sea-introduced.
(10)
68
WILLIS AND GARDINER : BOTANY
ROSACEÆ.
Rosa, L.
I. Didi, 81, is a species of Rose ! Native name, Fini-fen-
ma. In Malé and Miladumadulu by mosques. — J. S. G.
Grassulacece.
Bryophyllum, Salish.
calycinum, Salish. ; Fatunfaifila (D.) ; Bodu faru (G.)
M. ; Akkapana, S.
I. Didi, 14 ! Malé, Haly, 1892, Gardiner ! Hulule, 18,
Gardiner ! “Not cultivated ; along paths near villages where
the path passes into jungle. Not on the beach. Very
sparsely through the group.” — J. S. G.
Cosmopolitan in the tropics. Not in the Laccadives or
Minikoi.
Rhizophoraceœ.
Rhizophora, L.
mucronata, Lam. ; Kadol, S. ; Mangrove.
Addu Atoll, Wiligili in Suvadiva Atoll, Gardiner ! “ In the
jungle of Maradu I. ; here there is no definite swamp, but
the whole is very dank. In damp jungle all over the group.”
— J. S. G.
Sea shores of the eastern tropics. Not recorded for the
Laccadives or Minikoi.
Bruguiera, Lam.
caryophylloides, Blume ; Kandu, M.
Goif urfehendu Atoll, Yeimandu in Kolumadulu, Gardiner!
Large mangrove swamps (in Maldivian, Kandufa = mangrove
or mangrove swamp) in Goidu in Goifurfehendu ; no
Rhizophora seen. Also in Miladumadulu. The young
rootlet when still green and hanging from the plant is very
generally eaten.
Sea shores, in mangrove swamps, not common, S. India,
Ceylon, Malaya. Not recorded for the Laccadives or
Minikoi.
OF THE MALDIVE ISLANDS.
69
Gomhretaceœ.
Terminalia, L,
Catappa, L. ; Midili (Didi) ; Medili (in Nudun, G.) ;
Dommadu (in Hulule, G.) ; Gobu (in Had-
dumati), M. ; Kottamba, S. ; the Country Almond.
I. Didi, 22 ! Malé, Christopher, 1888 ! Hulule 5, Vei-
mandu in Kolumadulu, Gardiner ! Also in Goifurfehendu
Atoll, Suvadiva, Addu, Gardiner. Abundant all over the
group, especially in dense high jungle. Always found on
beaches growing outwards by sand additions, a little back
from the beach proper. Seeds much eaten in Suvadiva and
Addu, where the tree attains a great size, and is one of the
most important constituents of the jungle.
A littoral species ; not known wild in India or Ceylon,
though often planted. Common in the Andamans, Nicobars,
Malaya, Mauritius, &c. Minikoi, Prain ; Diego Garcia,
Hemsley. * It is probably native in the Maldives, introduced
by the sea.
Lumnitzera, Willd.
racemosa, Willd.
Gaddu in Suvadiva, Gardiner ! Also in Wiligili, Suva-
diva.— J. S. G. I. Didi, 12. May be this, or may be L.
coccinea.
Mangrove swamps of the eastern tropics. Ceylon. Not
recorded for the Laccadives or Minikoi.
coccinea, W.& A.
I. Didi, 12. Native name Kandu (c/. other mangroves) is
very probably this sp. Found in mangrove swamps, Ceylon,
Burmah to Polynesia.
Gyrocarpus, Jacg[.
Jacquini, Uoxh,
Maldives, Gardiner (Prain) ! One of our specimens was
identified as this species by Dr. Prain, but has unfortunately
become mislaid. “ The specimen was no doubt from Hedu-
f uri, where it occurs on the sandy part of the island. It is
70
WILLIS AND GARDINER : BOTANY
not an uncommon bushy tree on the same area through the
whole group, often growing on the driest part of a sand flat.”
— J. S. G.
Cosmopolitan in the tropics. The wood is used for out-
riggers and catamarans in Ceylon, and perhaps the tree may
have been intentionally introduced, but it is more probably
sea-borne.
Myrtacece.
PSIDIUM, L.
Guyava, L. ; Féru, M. ; Fera, S. ; the Guava.
I. Didi, 83 ! Trimen’s list ; no specimen seen. In Malé,
Gardiner.
Cultivated in all hot countries. Cultivated in Minikoi,
but not in the Laccadives.
Eugenia, L.
Malaccensis, L.; Jumbu, M. the Malay Apple.
I. Didi, 50, a fragment, appears to be this. Fruiting in
Male, and a few young trees in Addu. — J. S. G.
A Malayan species, commonly cultivated in India and
Ceylon.
Jambos, L. ; Jumbu, M. ? Jambu, S. ; the Rose
Apple.
Trimen’s list ; no specimen seen. Male and Addu. —
J. S. G.
Malayan ; cultivated in India and Ceylon. Minikoi, cult.,
Prain.
Javanica, LamA;. Jamburool, M.
I. Didi, 53, a fragment, is perhaps this species.
Jambolana, Lamk.; Lami, M.; Ma-dan, S.; Naval, T. ;
the Black Plum.
I. Didi, 19 !
Common in Tropical Asia, Ceylon, &c. Probably intro-
duced for its fruit. Cultivated in Minikoi.
Barriiigtoiiia, Forst
speciosa, Forst; Kimbi (G.), M. ; Mudilla, S.
OF THE MALDIVB ISLANDS.
71
I. Didi, 9 ! Hulule 6, Gardiner ! Two or three trees in
the S.W. of Hulule ; rare in the group, especially in the
‘ north. Heebahdu in the north of Suvadiva Atoll covered
with it ; four or five trees in Midu in Addu Atoll. Also on
Turahdu in S. Mahlos. — J. S. G.
Sea shores, Ceylon, S.W., very rare, but now often planted.
Andamans to Polynesia, a littoral species, probably intro-
duced by the sea. Not recorded for Minikoi or the Lacca-
dives. Diego Garcia, Hemsley.
Lythracece.
(Ammaniabaccifera, L. ; in Kalpeni, Laccadives, Prain.)
Pemphis, Forst.
acidiila, Forst.; Dhadukuradi, Kuredi, Kuradu, M.
I. Didi, 28 ! Male, Christopher, 1888 ! Goifurfehendu
Atoll, Hulule 4, Veimandu in Kolumadulu, Gardiner ! Also
in Turahdu, South Mahlos, Gardiner. All along the western
beach of Holule, a narrow belt one shrub thick, the roots
washed by every high tide. The shore washes away, and
bushes may be seen growing right out in the water. Com-
mon in all the Maldives. Infested with Sphinxes and small
moths at night ; it is useless to try any other plant for them
if there be Pemphis near. — J. S. G.
A littoral species, common on the shores of the Old World
tropics.
Minikoi, Gardiner. Not recorded for the Laccadives.
Lawsonia, L.
Alba, Lam. ; Heena (Innapa, misprinted
Junapa in Trimen’s list), M. ; tlie Henna.
I. Didi, 130 ! Trimen ; no specimen, Pyrard de Laval.
Goifurfehendu Atoll, Gardiner ! “ In cultivated land, scarce
in the Archipelago.” — J. S.*G.
Ceylon and India, frequently cultivated. Minikoi, culti-
vated, Prain.
Sonneratia, Linn.f.
acida, Linn, f.; Kulowa, M. ; Kirilla, S.
72
WILLIS AND GARDINER : BOTANY
I. Didi, 30 ! Cultivated in Landu, Male, and Addu,
Gardiner. Trimen’s list ; no specimens.
With the mangroves, Ceylon Bengal to Java. Not
recorded for Minikoi or the Laccadives. Most likely intro-
duced into the Maldives, but possibly brought by the sea.
PUNICA, L.
Granatum, L. ; Annaru, M. ; Anar, Hind. ; Delun,
S. ; the Pomegranate.
I. Didi, 115 ! Trimen’s list (mentioned by Ibn Batuta).
Fainuin North Mahlos,Turahdu in South Mahlos, Veimandu
in Kolumadulu, Gardiner ! Cultivated by the villages in
many parts of the Archipelago.
Cultivated in all tropical and subtropical countries. Mini-
koi and Laccadives (Ameni, Akati), Prain.
PassifloracecB.
Passiflora, L.
COERULEA, L.
Male, Gardiner ! Only seen in Malé. — J. S, G.
An American species ; cultivated in Ceylon for its pretty
flowers.
[suberosa, L. ; in Diego Garcia, Hemsley.]
Carica, L.
Papaya, L. ; Falo, M. ; Papaiya, Hind. ; the Papaw,
I. Didi, 1 ! Turahdu in South Mahlos, Goifurfehendu,
Hulule, &c. — J. S. G.
“Three kinds are cultivated, one with long large fruits on
the main stem (falo), one with round small fruits on the
main stem (kuda falo), and a third (veo falo) with many
fruits on pendulous shoots. In Horsburgh Atoll the fruit
only reaches a small size.” — J. S. G.
Cultivated in all tropical countries. Minikoi, Laccadives,
in Kiltan, Anderut, Akati, and Kadamum, Prain.
OF THE MALDIVE ISLANDS.
73
CUCURBITACEÆ.^:*
Trichosanthes, L.
CUCUMERINA, L.
Fainu in North Mahlos, Gardiner! Rare, and in Addu.
Cultivated and wild in India, Malaya, North Australia.
Almost certainly introduced intentionally.
Benincasa, Savi,
CERIFBRA, Savi.; Fufu, M. ; the Ash Pumpkin.
I. Didi, 102. A fragment, is probably this species. Culti-
vated in the eastern tropics. “ Not seen.” — J. S. G.
Momordica, L.
CHARANTIA, L, ; Faga, M. ; Karivila, S. ; Pakal, T.
I. Didi, 61 !
Cosmopolitan in the tropics. Cultivated in Minikoi.
Cucumis, L,
SATIVUS, L. ; Kekuri, M. ; Kekiri, S. ; the Cucumber.
Hedufuri, Hulule 12, Yeimandu in Kolumadulu, Gardiner.
“ Extensively cultivated in Miladumadulu and the north,
not seen in the south.” — J. S. G.
Cultivated in all hot countries. Minikoi, Gardiner; Lacca-
dives, in Kiltan, Prain.
(Melo, L. ; the Melon, in Kadamum, Kiltan, and Akati
(Laccadives), and in Minikoi, Prain. Perhaps mistaken for
the preceding in some of the Maldives.)
(Luffa ægyptiaca, Mill. Cultivated in Minikoi, Prain.
I. Didi, 42., a mere fragment, native name Tora, may
possibly be this.)
CiTRULLUS, Schrad.
COLOCYNTHIS, Schrad. ?
I. Didi, 48, native name Kara, may perhaps be this,
VULGARIS, Schrad.; the Water Melon.
* A considerable variety of Cucurbitaceæ have in recent years been
introduced to Malé, partly for ornamental purposes and partly as
possible food. The soil is too poor in most islands for them, and they are
generally pronounced to be no use for eating. — J. S. G.
74
WILLIS AND GARDINER : BOTANY
Commonly cultivated in the Archipelago.
Cucurbita, L.
MOSCHATA, Duch, ; the Musk Melon.
Trimen’s list ; no specimens. Maradu and other islands
in Addu Atoll. — J. S.G. Cultivated in warm countries.
? Pepo, L. ; the Pumpkin ; Buruho, M.
Two sorts of pumpkin with yellow flowers are cultivated,
and known as Burubo in Hulule and elsewhere. — J. S. G.
I. Didi, 78. With native name Kadu, is probably some
kind of Cucurbita.
? MAXIMA, Duchesne ; the Gourd or Giant Pumpkin.
This is cultivated in Minikoi (Prain), and mentioned, but
without confirmatory specimen, in Trimen’s list, as seen by
Mr. Bell. The remarks just made about C. Pepo may refer to
this species. On trellises in Addu (? this species). — J. S. G.
[Cephalandra indica, Naud. Occurs in Akati, Laccadives,
Prain.]
Ficoidece.
Sesuvium, L.
Portulacastrum, L. ; Mapijja, M.
Hulule, F. Lewis !
Cosmopolitan tropical littoral. Minikoi, Prain.
ARALIACEÆ.
Panax, SP. ? A fragment from Malé, Gardiner, growing
near mosques, appears to be a Panax. Probably cultivated
for ornament.
Arabia, L.
Guilpoylei, F. von M.
Kaddu, in Haddumati Atoll, Gardiner ! Evidently in-
troduced from Ceylon as an ornamental plant.
GAMOPETALÆ.
Ruhiacece,
(Dentella repeiis, Forst. Occurs in Anderut, Laccadives.)
OF THE MALDIVE ISLANDS.
75
OMenlandia, L.
corymbosa, LJ
Male, 1892, Haly. A somewhat doubtful fragment, named
by Trimen. A cosmopolitan tropical weed. Laccadives, in
Kalpeni, Kadamum, Kiltan, Prain.
(diffusa, Boxh. ; in Anderut (Laccadives) and Minikoi,
Prain.)
umbellata, L.; Emmuli, M. ; Saya, S.; Chaya, T. ; Chay
Root.
I. Didi, 135 ! Hedufuri, on cleared grain land, not
common. Goifurfehendu Atoll, Gardiner ! Trimen’s list.
A weed of sandy ground, especially near the coast,
Ceylon, India, Burmah. The root yields a dye, formerly
much used in Ceylon. “ May have been introduced as dye-
plant, but certainly not now used as such.” — J. S. G.
biflora, L.; Beem magu (Hulule), Eyaganawatura,
Kudingaybelamaw (Minikoi), M.
Trimen’s list ; no specimen. Goifurfehendu Atoll, Hedu-
furi, on cleared grain land, not common, Hulule 52, in dry
sandy paths, Veimandu in Kolumadulu, Gardiner ! “ In
suitable places throughout the group.” — J. S. G.
Common near the coast in Ceylon and Tropical Asia.
Laccadives, in Kalpeni and Kadamum, Prain ; Minikoi,
Gardiner.
H AMELIA, Jacq.
Patens, Sw.
Malé, F. Lewis.
S. America ; cultivated in Ceylon.
Guettarda, L.
speciosa, L.; Uni, M. ; Nil-pitcha, S. ; Panir, T.
I. Didi, 166 ! Pyrard de Laval. Hedufuri, Hulule 61.
Veimandu in Kolumadulu, Gardiner ! Very common
throughout the group ; occupies any position not upon the
beach. Sweet-scented flowers, occasionally used.
(11)
76
WILLIS AND GARDINER : BOTANY
A cosmopolitan tropical littoral species ; Ceylon only on
the S.W. coast ; Laccadives, in Kadamum, Prain ; Minikoi,
Gardiner ; Diego Garcia, Hemsley.
(IXORA COCCINBA, L. ; in Kadamum, Anderut, Kalpeni
(Laccadives), and Minikoi, Prain.)
florinda, L.
citrifolia, A., var bracteata, Hk, /. ; Ahi, M. ; Ahu, S.
I. Didi, 148 ! Goifurfehendu Atoll, Hulule 38, Veimandu
in Kolumadulu, Gardiner !
“ This species and Ochrosia borbonica very quickly make
their way to any fresh sandbanks that form— a common
occurrence — and these two alone very often form a belt
separating the cocoanuts from the sea. Scævola commonly
grows between them again and the water.” — J. S. G.
“ Very numerous in Hedufuri, to some extent planted by
the people for the root, from which a permanent red dye for
cloth is obtained by mixing a decoction with lime.” — J. S. G.
A littoral species, common in the tropics of Asia and
Australia. Ceylon, from Colombo to Tangalla. Laccadives,
in Bangaro, Kalpeni, Kiltan, Akati, and Kadamum, Prain ;
Minikoi, Gardiner, Prain ; Diego Garcia, Hemsley.
Spermacoce, L,
ocyinoides, Burm.f.
Hedufuri, Gardiner ! Near mosques in cleared land in
many islands, but not abundant.
A common weed in Tropical Asia and ilustralia, rare in
Ceylon. Not recorded for the Laccadives of Minikoi.
Compositce.
Vernonia, Schrei,
cinerea. Less. ; Walu Kafa, M. (Kambulichi, Minikoi);
Monarakudimbiya, S.
Hedufuri, Goifurfehendu Atoll, Hulule 51, 68, Veimandu
in Kolumadulu, Gardiner ! Common in cleared land
throughout the group.
OF THE MALDIVE ISLANDS.
77
A common palæotropical weed. Laccadives, in Ameni,
Anderut, Kadamum, Kiltan, and Akati, Prain ; Minikoi,
Gardiner, Prain ; Diego Garcia, Hemsley.
Adenostemma, Forst.
viscosum, Forst. ; Foni-loli, M.
Hedufuri, Hulule 58, Veimandu in Kolumadulu, Goifur-
fehendu Atoll, Gardiner !
I. Didi, 16, 121 ! Common in the north and very com-
mon in Addu ; in the lower part of the jungle.
A cosmopolitan tropical weed. Laccadives, in Kalpeni,
Prain ; Minikoi, Prain.
Agératum, L.
conyzoides, L. ; Kochché-fai, M. ; Hulantalla, S. ;
Pumpullu, T.
I. Didi, 73 ! Malé, Christopher, 1888, Gardiner ! Kaddu
in Haddumati Atoll, Gardiner !
A cosmopolitan tropical weed, probably of American
origin. Laccadives, in Ameni, Anderut, Kalpeni, and Kiltan;
Minikoi, Prain ; Diego Garcia, Hemsley.
Blumea, DC.
(laciiliata, DG.; in Kiltan and Akati, Laccadives,
Prain.)
membraiiacea, DG.
Veimandu in Kolumadulu,Kaddu in Haddumati, Gardiner!
The forms collected closely resemble those found in Ceylon.
A common Asiatic tropical weed.
Eclipta, L.
alba, Hassk. ; Kalukadili, M. ; Kikirindi, S.
I. Didi, 62 ! Veimandu in Kolumadulu, Gardiner !
A cosmopolitan tropical weed. Laccadives, in Ka'damum
and Kiltan, Prain.
Wedelia, Jacq.
(caleiidulacea, Less.; in Anderut, Laccadives, Prain.)
biflora, DG. ; Mirihi, M.
n
78
WILLIS AND GARDINER : BOTANY
I. Didi, 43 ! Hedufuri, Goifurfehendu Atoll, Hulule 57,
62, Veimandu in Kolumadulu, Gardiner ! Common in the
entire group.
A littoral species, common from Ceylon to Singapore.
Laccadives, in Kadamum, Ander ut, and Kiltan, Prain ; Mini-
koi, Prain, Gardiner.
(Bidens pilosa, L, ; in Kadamum, Laccadives, Prain.)
(Crépis acaulis, Hh, /. ; in Kiltan, Laccadives, Prain.)
Tithonia, Desf.
diversifolia, A. Gray ; Bodu-mirihi, M.
I. Didi, 47 !
A native of the Southern United States and Mexico, now
very common in Ceylon ; probably intentionally introduced
into the Maldives. “ Not seen.” — J. S. G.
Artemisia, A.
vulgaris, L. ; Mirajjé Kochchefai, M. ; Wal-kolundu, S.
I. Didi, 34 !
A weed of Europe, temperate Asia, Siam, and Java ; an
escape in Ceylon. Probably intentionally introduced for its
fragrant leaves.
Emilia, Gass.
soncliifolia, DC.; Hirikulla, M. ; Kadupara, S.
I. Didi, 124 ! Goifurfehendu Atoll, Gardiner ! “ Common
in the south of the Archipelago.” — J. S. G.
A common weed of the Old World tropics.
Lactuca, L.
polycepliala, Benth.
Goifurfehendu Atoll, Gardiner. A somewhat abnormal
specimen is this species in all likelihood, according to Dr.
Prain. It is a native of North India.
Laiinæa, Gass.
pinnatifida, Gass. ; Dandu filia ; Kulla fila, Kadapi
(Minikoi), M.
I. Didi, 18 ! Goifurfehendu Atoll, Hulule 56, Gai diner!
North Mahlos.— J. S. G.
OP THE MALDIVE ISLANDS.
79
“ One of the first to appear on a bare sandbank, and the
outermost on a growing sandy beach.” — J. S. G.
A littoral species, common on the coasts of Ceylon, India,
Egypt, Mauritius, and E. Africa. Laccadives, in Bitrapar and
Kadamum, and in Minikoi, Prain, Gardiner.
Goodeniacece,
Scævola, L.
Koeiiigii, Vahl.; Magu, M. ; Takkada, S.
Hedufuri, Goifurfehendu Atoll, Hulule 49, Veimandu in
Kolumadulu, Gardiner !
I. Didi, 142 ! Trimen’s list ; no specimen. The most
numerous shrub in the group ; wherever a sandy beach is
growing outwards it lines the shore. (See under Pemphis,
Morinda, Ochrosia, and below.)
A littoral species, common in E. Tropical Africa, Asia, and
Australia, and in Polynesia. Common in Ceylon, especially
in the S.W.; Laccadives, in Bitrapar, Kadamum, Kiltan, and
Akati, Prain ; Minikoi, Prain, Gardiner ; Diego Garcia
Hemsley.
{Plumbaginaceœ,
Plumbago zeylanica, L. ; in Ameni, Laccadives.)
Myrsinaceœ.
Irdisia, Sw,
humîlîs, Vahl,
Suvadiva Atoll, very abundant ; the undergrowth mostly
composed of it ; Addu Atoll, common in all the jungles,
Gardiner (C. Mez ) !
Ceylon S.W. and Batticaloa, India to China. The Ceylon
distribution would seem to indicate sea-carriage, but it may
have been introduced by birds.
(The absence of Ægiceras majus, Gaertn.^ one of the
palæotropical mangroves, from the Maldives, as well as from
the Laccadives and Minikoi, is noteworthy. It may however
very likely have been overlooked.)
80
WILLIS AND GARDINER : BOTANY
Sapotaceœ.
Chrysophyllum sp. ?
I. Didi, 45, native name Sabudeli, is probably a sp. of
Chrysophyllum.
Ulimusops, L.
Elengi, L. ; Munima, M. ; Munamal, S.
I. Didi, 116 !
Ceylon, India, Malaya, often, cultivated, and probably
introduced into the Maldives intentionally, for the sake of
its timber.
OLEACEÆ.
Jasminum, L.
Same AC, Ait.; Irudema, Re Irudema, M.; Pichcha, S.;
Arabian Jasmine.
1. Didi, 87 ! Pyrard de Laval (onudemaus). Hedufuri,
Goifurfehendu Atoll, Hulule 25, 35, Veimandu in Koluma-
dulu, Kaddu in Haddumati, Addu Atoll, common in the
jungle of the whole atoll, Gardiner! Male, Christopher, 1 888 !
An Indian species, commonly cultivated in Ceylon and
India. Most likely or certainly introduced for cultivation,
but now more or less wild.
Auriculatum, Yalil.; Kudima, M.
I. Didi, 46 !
Common in the dry parts of Ceylon and in India.
Grandiflorum, L.; Huwanduma, M.
I. Didi, 80 ! Hedufuri, Goifurfehendu Atoll, Hulule 24,
Veimandu in Kolumadulu, Gardiner ! “ Cultivated round
mosques and holy men’s graves. The flowers are soaked in
fine cocoanut oil, to which a trace of ambergris is added. With
this mixture the natives rub themselves after washing tj
scent themselves before going into the mosques.” — J. S. G.
A native of N.W. India, cultivated in India and Ceylon,
and evidently intentionally introduced.
OF THE MALDIVE ISLANDS.
81
Apocynacece,
Thevetia neriifolia, Juss. ?
I. Didi, 66, native name Hinbatu, is almost certainly this
species, which is cultivated in Ceylon as an ornamental plant.
Ochrosia, ÄSS.
borboilica, Gmel.; Dhumburi , Dumburi, M.; Mudu-
kaduru, S.
I. Didi, 27 ! Male, Capt. Simons ! Hedufuri, Hulule 43,
Gardiner ! This and Morinda citrifolia very quickly make
their way to any fresh sandbanks that form, and these two
alone often form a belt separating the cocoanuts from the
sea with Scævola between them and the water.
A littoral species, Mascarene Is., Ceylon on the S.W.
coast, Andamans, Nicobars, Singapore, Java. Not in India
or the Laccadives. Minikoi, Prain; Diego Garcia, Hemsley.
Vinca, L.
rosea, L.; Maliku ruva, M.
Male, Haly, 1892. Christopher, 1888. Gardiner !
I. Didi, 170 ! Cultivated almost everywhere,
A cosmopolitan tropical weed. Abundant on the S.W.
coast of Ceylon.
Plumeria, L.
ACUTIFOLIA, Poir.; Gulu sampa. Semper Beddha, M.;
the Temple Tree.
I. Didi, 26 ! Veimandu in Kolumadulu, Gardiner ! Tri-
men’s list, no specimen. Cultivated and semi-wild in Ceylon.
Probably introduced intentionally for its scented flowers,
which in Ceylon are very largely used as temple offerings
by the Buddhists. Two or three trees by every mosque in
the group. The flowers used like those of Jasminum grandi-
fiorum, q.v.
Asclepiadaceœ.
Calotropis, Br.
gigailtea, Br. ; Ruva, Hudu ruva, Rua, M. ; Wara, S.
82
WILLIS AND GARDINER : BOTANY
Male, Haly, 1892 ! Hulule 36, 78, Gardiner ! I. Didi, 49
and 74, fragments, are almost certainly this species. Culti-
vated by every large village. The milk is used for wounds,
and rubbed with a decoction of tea under the eyes for sun. —
J. S. G. The stem fibre used for fishing lines, and the fioss
of the seeds for charms in magic (J. J. Thorburn).
Common in Ceylon, India, Malaya, S. China, Laccadives,
in Kadamum, Prain . Probably intentionally introduced into
the Maldives.
(Asclepias curassavica, L. ; in Diego Garcia, Hemsley.)
(Tylopliora asthmatica, W.& A.; in Kadamum, An derut,
Kiltan, and Akati, Laccadives, Prain.)
(Leptadeiiia reticulata, W. â A.; in Kadamum and Kiltan,
Laccadives, Prain.)
Boraginacece,
Cordia, L,
subcordata, Lam,; Kani, M.
I. Didi, 165 ! Goifurfehendu Atoll, Hulule 17, Vei-
mandu in Kolumadulu, Gardiner ! In Suvadiva Atoll. —
J. S. G. A regular constituent of dense jungle wherever
it is found in the Maldives.
A littoral species of the eastern tropics. Ceylon only at
Trincomalee. Diego Garcia, Hemsley. Not recorded for
Minikoi or the Laccadives.
I. Didi, 55, native name Ragunda, a mere skeleton of a
leaf, is a Cordia, perhaps another sp.
Tournefortia, L.
argeiitea, L.f.; Bori, Mabori, M. ; Karan, S.
I. Didi, 107 ! Heduf uri, Goifurfehendu Atoll, Hulule 60,
Veimandu in Kolumadulu, Gardiner ! Male, Christopher,
1888 ! Also in N. Mahlos. — J. S. G. Contests the shore
with Scævola in the same positions — the 'only plant which
does so. Also with Pemphis, but its roots cannot stand
the direct tidal action like those of that genus.
OF THE MALDIVB ISLANDS.
83
A littoral species, Mauritius, Ceylon, rare on the S.W.
coast and at Trincomalee, not in India ; Nicobars, Malaya,
Australia, Andamans, Minikoi, Prain ; Laccadives, in Bitra-
par, Kiltan, Kadamum, Akati (seedlings germinating on the
shore), Prain ; Diego Garcia, Hemsley.
Trichodesma, Br,
zeylaiiicum, Br.; Ma-lebu, M.
I. Didi, 172 ! Fainu, in N. Mahlos, on land gone out of
cultivation, Goifurfehendu Atoll, Addu Atoll, Gardiner !
A common weed, Mascarene Is. to Australia.
, Gonvolvulacece,
Ipomœa, L.
gran diflora, Lam.
Goifurfehendu Atoll, Addu Atoll (common in the jungle
everywhere), Gardiner ! In the jungle on rocky land, and
creeping over rocky beaches being washed up.
A littoral species, E. Africa and Mascarene Is. to Polyne-
sia, Diego Garcia, Hemsley ; Laccadives, in Bitrapar
(densely draped over the clumps of Scævola and Tourne-
fortia) and Kadamum, Prain.
Batatas, Lam. ; Oludukattala, M. ; Battala, S. ;
Sweet Potato.
I. Didi, 75 ! Hulule, scarce, Gardiner! Grown here and
there throughout the whole group, especially in Addu.
Cultivated in all warm countries. Laccadives, in Ameni,
Ander ut, Akati.
Quamoclit, L.; Kudiraima-veyo, M.
I. Didi, 33 !
An American species, naturalized in Ceylon, with pretty
flowers.
Turpethum, Br. ; Kurifila, M.
I. Didi, 23 !
Cultivated and wild in the tropics of the Old World.
Perhaps an accidental, but more probably an intentional,
introduction.
(12)
84
WILLIS AND GARDINER : BOTANY
deiitieiilata, Chois.
Veimandu in Kolumadulu, Gardiner !
A littoral sea-borne plant, Seychelles to Polynesia ; not in
India. S.W. Ceylon (rare). Minikoi, Prain.
biloba, Forsk. (Pes-capræ, Roth.); Taburu, Thaburu,
M. ; Mudu-bin-tamburu, S.
I. Didi, 41 ! Goifurfehendu Atoll, Gardiner ! Very
uncommon, not seen elsewhere ; common on the lagoon
side of Goidu. — J. S. G.
Cosmopolitan on tropical shores. Laccadives, in Bitrapar,
Kalpeni, Akati, and Kadamum, Prain ; Minikoi, Prain,
Gardiner ; Diego Garcia, Hemsley.
(siniiata, Ortega; in Minikoi, Prain, Gardiner.)
Hewittia, W. & A.
bicolor, W. & A.
Hedufuri, Gardiner ! Only seen in Mahlos. — J. S. G.
Ceylon, S.E. India, Malaya, Tropical Africa. Probably
unintentionally introduced with cultivated plants.
Convolvulus, L.
parviflorus, Vahl. ; Walu mirihi, M.
Hedufuri, Goifurfehendu Atoll, Hulule 32, 53, Veimandu
in Kolumadulu, Gardiner ! Kitchen middens throughout
the group. — J. S. G.
Ceylon, rare, S. India, Malaya, Australia, Tropical Africa,
Minikoi, Prain. Probably introduced by the sea, but
perhaps introduced with cultivated plants.
Evolvulus, L.
alsinoides, L. ; Veo magu, Meia limbo, M.
Hedufuri, Hulule 50, 64, Kaddu in Haddumati Atoll,
Gardiner ! On sandy paths which are periodically cleared.
In the N. chiefly. — J. S. G.
A weed of dry places, cosmopolitan in warm countries,
liaccadives, in Kadamum.
OF THE M ALDI VE ISLANDS.
85
Solanaceœ.
Solanum, L.
Mblongena, L.; Bari, Kara, M. ; Wambatu, S. ; the
Brinjal.
I. Didi, 112 ! Trimen’s list. Veimandu in Kolumadulu,
Gardiner ! Cultivated in the whole group, very abundant
in Addu.— J. S. G.
Cultivated in all warm countries. Minikoi, Gardiner.
Not recorded for the Laccadives.
(torviim, Siv. ; in Minikoi, Brain.)
Physalis, L.
mîiiîmv), L.; Muraki, M.; Mottu, S.
I. Didi, 94 ! Malé, Christopher, 1888 ! Goifurfehendu
Atoll, Hulule 11, Veimandu in Kolumadulu, Gardiner !
Sparsely by villages in the whole group, the fruit eaten in
curry.
A common weed in the tropics of the Old World. Perhaps
introduced by birds or by the sea. Laccadives, in Bangaro,
Anderut, Kadamum, Kiltan ; Minikoi, Brain.
(Peruviana, L. ; cultivated in Ameni, Laccadives, Brain.)
Capsicum, L,
MINIMUM, Roxh.; Mirus, M. ; Nayi-miris, S. ; Bird
Pepper.
I. Didi, 106 ! Horsburgh Atoll, Male, Veimandu in Kolu-
madulu, Gardiner ! Mahlos, Miladumadulu, Gardiner !
Everywhere cultivated.
Cultivated in all hot countries. Laccadives, in Akati,
Brain.
(Frutescens, L. ; cult. in Minikoi, Brain.)
Datura, L.
fastuosa, L. ; Orhani, M. ; Attana, S.
I. Didi, .132 ! Landu in Miladumadulu, Veimandu in
Kolumadulu, Gardiner ! Also in Addu Atoll. — J. S. G.
A cosmopolitan tropical weed, possibly introduced by
birds or with cultivated plants. Laccadives, in Ameni,
Anderut, Akati, Kiltan, Brain ; Minikoi, Gardiner, Brain.
86
WILLIS AND GARDINER : BOTANY
SUAVEOLENS, H, & B.
Male only, Gardiner ! Cultivated in Ceylon, &c., round
native houses.
Nicotiana, L.
Tabacum, L. ; Dumpai, M. ; Dun-kola, S. ; Tobacco.
Many plants in Male, in house enclosures, and a few in
Mahlos and Miladumadulu. — J. S. G.
Everywhere cultivated. Not in Minikoi or the Lacca-
dives.
ScropJmlariacece.
(Liiiaria ramosissima, Wall. ; in Kiltan, Prain.)
Herpestis, Gaertn.f.
Monniera, H. B. & K.; Veppila, M. ; Lunu-wila, S.
I Didi, 145 ! Landu, in Miladumadulu, in yam patches,
Gardiner !
A cosmopolitan tropical weed. Common in marshy places
near the coast in Ceylon. Laccadives, in Anderut, Prain ;
Diego Garcia, Hemsley.
[Striga liitea. Lour.; in Akati, Kadamum, Kiltan, Lacca-
dives, Prain ; Diego Garcia, Hemsley.]
Acantliaceœ.
Ruellia, L.
ringeiis, L. (prostrata, Lam., var. dejecta); Nita bodi,
Nitu bad], M. ; Nilpuruk, S.
I. Didi, 5 ! Hedufuri, Horsburgh Atoll, Malé near
mosques. Hulule 76, Veimandu in Kolumadulu, Kaddu in
Haddumati, Gardiner Î Also in Mahlos in many islands.
— J. S. G.
A common weed in India, Ceylon, N. Africa. Minikoi,
Prain.
Barleria, L.
Prionitis, L. ; Ma tumba, Hai kurudo (Minikoi), Dai
kurandu, M. ; Katu-karandu, S.
OF THE MALDIVE ISLANDS.
87
I. Didi, 173 ! Mafaro in Miladumadulu, Hulule 77, Goi-
furfehendu Atoll, Veimandu in Kolumadulu, Gardiner !
Cultivated by villagers every\vliere. — J. S. G. Used in
native medicine (J. J. Thorburn).
Tropical Africa and Asia, often cultivated for medicinal
uses, and perhaps or probably intentionally introduced into
the Maldives. Laccadives in Ameni ; Minikoi, Gardiner,
Prain (planted as a hedge).
Justicia, L.
procumbeiis, A.
Trimen’s list ; no specimens.
A common weed in Tropical Asia and Australia.
Rungia, Nees.
parviflora, Nees.
Hulule 55, Gardiner ! “ Very common on the roots of the
cocoanut trees ; frequent in the group (Addu, Landu in
Miladumadulu, &c.)” — J. S. G.
A common weed of Tropical Asia. Laccadives, in Ameni,
Akati, Kadamum, Prain.
(lillifolia, Nees ; in Kadamum and Akati, Laccadives,
Prain.)
(Peristrophe hicalyculata, Nees; in Ameni, Kadamum,
Kiltan, Laccadives.)
Verbenaceœ.
Laiitaiia, L.
mixta, L.
Hedufuri, Gardiner ! Common in cultivated land all
over the north of the Archipelago. — J. S. G.
An abundant weed in Ceylon, S. India, &c. Not recorded
for the Laccadives or Minikoi. Probably carried by birds
in the first place.
Lippia, L.
iiodiflora, Rich. ; Hunigunditila, M. ; Heri-mena-
detta, S.
88
WILLIS AND GARDINER : BOTANY
I. Didi, 162 ! Malé, Trimen ; no specimen. Veimandu in
Kolumadulu, Addu Atoll, Gardiner ! A very common yam
garden and jungle weed all over Addu. — J. S. G.
A cosmopolitan tropical weed, especially near the sea in
Ceylon. Perhaps introduced by birds (Prain). Minikoi,
Gardiner, Prain ; Laccadives, in Akati, Prain ; Diego Garcia,
Hemsley.
Stachytarpheta, Vahl.
indica, Vahl., var. jaiiiaicensis ; Malaembu, Nunnay
(Minikoi), M. ; Bala-nakuta, S. ; Nai-oringi, T.
I. Didi, 136 !
A common tropical weed, possibly intentionally intro-
duced. Minikoi, Gardiner; Laccadives, in Anderut, Prain.
Tectona, L.f.
Grandis, L.; Teak.
Malé, cultivated, Capt. Simons !
Premiia, L.
integrifolia, L.; Kude, Gelavalie (both Minikoi), M.;
Midi, S.
Goif urfehendu Atoll, Gardiner ! Common in the densest
jungle, especially in Suvadiva. On rocky ground, one of
the first bushes to appear on a rocky bank. Very common
in N. Mahlos.
A littoral and inland species, common in Ceylon, India,
the Malay Peninsula, Andamans, Nicobars, Laccadives, in
Kadamum and Kalpeni, Prain ; Minikoi, Gardiner, Prain ;
Diego Garcia, Hemsley,
Vitex, L.
Negundo, L. ; Dunnika, M. ; Nika, S. ; Vennochi, T.
I. Didi, 123 ! Turahdu in S. Mahlos, Veimandu in
Kolumadulu, Gardiner ! Male, Christopher, 1888 ! Used in
medicinal baths (J. J. Thorburn).
Common in Tropical Asia. Used medicinally in Ceylon
and possibly intentionally introduced into the Maldives.
OF THE MALDIVE ISLANDS.
89
Clerodendroii; L.
inerme, Gaertn.; Dugeti, Dugajde (Minikoi), M. ;
Wal-gurenda, S.
Malé, Haly, 1892! I. Didi, 118 !
A littoral species, abundant on the coasts of Ceylon, India,
and Burmah. Probably introduced by the sea. Minikoi,
Gardiner, Prain ; Laccadives, in Kalpeni, Prain.
[The absence from the Maldives, Laccadives, &c., of
Avicennia officinalis, A., one of the common mangroves, is
noteworthy.]
Lahiatce.
Ocimum, L.
basiliciim, L. ; Gandakoli (Minikoi), M. ; Suvanda-
tala, Tala-kola, S. ; the Sweet Basil.
I. Didi, 84! Fainu in N. Mahlos, Hedufuri, Goifurfe-
hendu Atoll, Gardiner !
Tropical Asia, Africa, America, often cultivated. Minikoi,
Gardiner.
gTatissiiiium, L. ; Fonitula, M. ; Otaka, S.
I. Didi, 134 !
Tropical Asia, Africa, America, naturalized in Ceylon.
Minikoi, Prain; Laccadives, in Bitrapar, Kadamum, Kalpeni,
Prain,
sanctum, L. ; Kulitula, Gai Kehabuli, M. ; Maduru-
tala, S. ; the Tulsi.
I, Didi, 59 ! Turahdu in N. Mahlos, Hulule 73, only by
villages, Gardiner !
Tropical Asia and Australia. Sacred to Hindus. Here
perhaps originally intentionally introduced ; now a weed.
Plectrantlms zeylanicus, Benth.?
Fragments, I. Didi, 125, native name Huvadukotun, and
Kaddu in Haddumati, Rimbudu in S. Nilandu (Huvaduka,
M.), Gardiner, are perhaps this species.
90
WILLIS AND GARDINER : BOTANY
Anisomeles, Br.
ovata, Br, ; Muskotan, Maskota (Minikoi), M. ;
Yakwanassa, S.
I. Didi, 90 ! Hedufuri, Goifurfehendu Atoll, Hulule 48,
Veimandu in Kolumadulu, Gardiner ! Also in Addu. —
J. S. G.
A common weed of Tropical Asia. Laccadives, in Kada-
mum and Akati, Minikoi, Gardiner, Prain,
Leucas, Br.
biflora, Veo miri, Veo mirihi, M. ; Getatumba,
S.
Hulule 54, 74, 78, only on damp paths scarce ; Wiligili in
Suvadiva, Gardiner ! “ On the roots of cocoanuts by the
kuna (Cyperus polystachyus) swamp in Wiligili. Common
also in the jungle of Maradu in Addu Atoll.” — J. S. G.
A common weed of Ceylon and S. India. Not in the
Laccadives or Minikoi.
zeylanica, Br.
Landu in Miladumadulu, in yam patches, Gardiner !
Also in other islands where yams are cultivated. — J. S. G.
A common weed of Tropical Asia. Not in the Laccadives
or Minikoi.
(aspera, Spreng,; in Ameni, Anderut, Akati, Kada-
mum, Kiltan, Laccadives, Prain.)
INGOMPLETÆ.
Nyctaginece.
Boerhaavia, L.
diffusa, L. (repens, L., var. diffusa, Hk. /.) ; Buran-
dagondi, Nanubedi (Minikoi), M. ; Pita-sudu-pala,
S.
I. Didi, 20 ! Goifurfehendu Atoll, Yeimandu in Kolu-
madulu, Gardiner ! Common in sandy sunny beaches
throughout the Archipelago.
OF THE MALDIVE ISLANDS.
91
A cosmopolitan tropical littoral species, introduced possi-
bly by the sea, but probably by birds. Laccadives, in
Bitrapar, Ander ut, and Kadamum, Prain ; Minikoi, Gardiner,
Prain ; Diego Garcia, Hemsley.
(repens, L. ; in Akati and Ameni, Laccadives, Prain.)
Mirabilis, L.
Jalapa, L.; Asuraffole re (red), Asuraffole hudu
(vrhite). Re asaruma, Asarumu, M.; Sendrikka,
S. ; Pattarashu, T. ; Marvel of Peru.
I. Didi, 71, 119, 114 ! Hulule 21, Gardiner ! Common
near villages, cultivated. Mentioned by Pyrard de Laval,
1602-7.
A native of Peru, cultivated , and escaped in all tropical
lands. Very common in Ceylon. Probably originally
introduced intentionally for cultivation. Laccadives in
Akati, Minikoi ; Prain.
Bougainvillæa, Comm,
SPECTABILIS, Wllld,
Malé, cultivated, Capt. Simons.
Pisonia, L.
aculeata. L. ?
A leafy bit from Veimandu, Gardiner, is probably this
species. Cosmopolitan in the tropics.
morîiidæfolia, Br. (alba. Span., probably = iiiermis,
Forst.) ; Los, M. ; the Lettuce Tree of Ceylon.
I. Didi, 7 ! Veimandu, Gardiner ! Male, east end. —
J. S. G
A littoral species, Andamans, Nicobars, Malayan, and
Polynesian Is. Laccadives, in Bitrapar, Prain ; Diego Garcia,
Hemsley. Probably sea-borne.
Amarantaceœ.
Celosia, L.
argentea, L,
Kaddu in Haddumati, Gardiner !
(13)
92
WILLIS AND GARDINER : BOTANY
Cosmopolitan tropical weed, often cultivated.
Amarantus, L,
spiuosus, L,
Malé, near mosques, Gardiner !
A cosmopolitan tropical weed.
Caudatus, L, ; Raidadisagu, M.
I. Didi !
Often cultivated in tropical and warm countries,
gaiigeticus, L,
Malé, Trimen ; no specimen. “ Not seen.” — J. S. G.
A cosmopolitan tropical weed, often used as a potherb.
viridis, L.; Sagu (Minikoi) ; Massagu, M. ; Kuru-
tampala, S. [Sag = potherb, Hind.]
I. Didi, 128 ! “ Not seen.” — J. S. G. Probably inten-
tionally introduced.
A cosmopolitan tropical weed, used as a potherb. Mini-
koi, Praip, Gardiner.
jVothosaerua, Wight.
brachiata, Wight
Malé, Haly, 1892 !
A weed of dry places and sea shores, Tropical Asia and
Africa.
Aerua, Forsk.
lanata, Juss, ; Hudufupila, M. ; Pol-kudu-pala, S.
I. Didi, 151 ! Malé, Christopher, 1888 ! Hedufuri, Goi-
furfehendu Atoll, Hulule 67, 72, Veimandu in Kolumadulu,
Gardiner ! Very common everywhere except in Addu.
A common weed, also littoral, in Tropical Africa and Asia
probably introduced by the sea (see notes on Atolls, below).
Laccadives, in Bitrapar, Kalpeni, Kadamum, Akati, Prain ;
Minikoi, Gardiner, Prain.
Achyranthes, L,
aspera, L, ; Karhi léibù, M. ; Gas-karal-heba, S.
OF THE MALDIVE ISLANDS.
93
I. Didi, 160 ! Hedufuri, Gardiner ! Only seen in Mahlos.
— J. S. G.
This specimen seems nearer to the type form than to the
littoral and inland var. porphyristachya. A cosmopolitan
tropical weed. Minikoi, Prain. The var. in Bitrapar, Ban*
garo, Kalpeni, Kiltan, Kadamum, Akati, and in Minikoi,
Prain ; Diego Garcia, Hemsley.
Polygonaceœ.
Antigonon, Endl,
Leptopus, H.& a.
Cultivated in Male, Capt. Simons.
(Polygonum barbatum, L. ; in Kalpeni, Laccadives,
Prain.)
PIPERACEÆ.
Piper, L.
Betle, L. ; Billi, M. ; Bulat-wel, S. ; Betel Pepper.
I. Didi 111 ! A few islands in Miladumadulu, and
all atolls. Plentiful in Addu. The leaf chewed. The leaf
is termed Wang, or in Addu Billaton. — J. S. G.
Everywhere cultivated in the tropics. Minikoi and Akati
Kiltan, Kadamum, Laccadives, Prain.
NIGRUM, L,; Pepper.
Trimen’s list ; no specimens.
A native of India, &c., everywhere cultivated in the tropics.
Not recorded for Minikoi or the Laccadives.
Lauracece.
Cassytha, L.
ûliforinis, L.; Dorn velam buli white sandhook).
Vele buli, M. ; Aga mula neti wela, S.
I. Didi 8 ! Hedufuri, Goifurfehendu Atoll, Hulule 31,
Veimandu in Kolumadulu, Gardiner ! Common in all the
islands where land is forming. Appears very early on a new
sandbank with the first shrubby vegetation, and dies out
when cultivation begins.
94
WILLIS AND GARDINER : BOTANY
Parasitic, rootless. Cosmopolitan on tropical shores. Lac-
cadives, in Kadamum and Kiltan, Prain. Not recorded for
Minikoi. Diego Garcia, Hemsley. Probably sea-borne in
the first instance and subsequently carried from island to
island by currents and birds.
Hernaiidia, L.
peltata, Meissn, ; Kando, Maskando, Kadu (Minikoi)
M. ; Palatu, S.
I, Didi 97 ! Hedufuri, Goifurfehendu, Hulule 46, Vei-
mandu in Kolumadulu, Gardiner ! The first large tree to
appear on a sandbank, where it occupies the very centre.
Fishing lines are made from the bark. Common in the
whole group.
A littoral species, Trop. E. Africa, Mascarene Is., Ceylon
S.W. coast, but not India, Andamans, Nicobars, Malay Penin-
sula and islands, N. Australia, Polynesia. Laccadives in
Korat, Prain ; Minikoi, Prain, Gardiner ; Diego Garcia,
Hemsley.
(ovigera, L. ; in Diego Garcia, Hemsley.)
Euphorhiaceœ.
Pbdilanthus, Neck.
TITHYMALOIDBS, Poit. ; Kala kiru, M.
Hulule 14, Gardiner ! Two houses at Male, bathing tank
at Hulule, and one house in Huludu, Addu Atoll. — J. S. G.
An American species, intentionally introduced for garden
use.
Euphorbia, L.
Atoto, Forst. ; Kiru tina, M.
Goifurfehendu Atoll, Hulule 19, Gardiner ! “ Frequent in
the group. Common in cultivated land near mosques.”“—
J. S.G.
A littoral sea-borne species, India (S.W. coast), Ceylon,
Malaya to Polynesia. Laccadives, in Ameni, Prain.
hypericifolia, L., var. parviflora, L.; Kerutina,
M. (Minikoi) ; Eladada-kiniya, S.
OF THE MALDIVE ISLANDS.
95
Hedufuri, in cleared grain land, not very common ; Vei-
mandu, in Kolumadulu, Addu Atoll, very common weed,
Gardiner !
A common tropical weed. Laccadives, in Kadamum and
Kiltan. Minikoi, Prain, Gardiner.
pilulifera, L. Qiirta^ L.) ; Kiriitina, M. ; Budada-
kiniya, S.
I. Didi, 104 ! Hedufuri, cleared grain land, not common,
Gardiner !
A cosmopolitan tropical weed. Laccadives, in Anderut,
Kadamum, Kiltan, and in Minikoi, Prain ; Diego Garcia,
Hemsley.
thymifolia, L.
Veimandu in Kolumadulu, Gardiner !
A weed of the tropics, except Australia. Minikoi, Prain.
Agyneia, VenU
bacciformis, A, Juss.
Addu Atoll, common in yam gardens and by villages,
Gardiner !
A littoral species, India, Ceylon, Java, Mauritius.
Phyllanthus, L.
(Emblica, L.; cultivated in Ameni, Laccadives,
Prain.)
maderaspatensis, L,; Meia limbo, Kudingke (Mini-
koi), M.
Hedufuri, Goifurfehendu Atoll, Hulule 63, Veimandu in
Kolumadulu, Gardiner ! Fairly common in the north.
A common weed of the tropics of Africa, Asia, Australia,
Minikoi, Prain, Gardiner. Laccadives, in Ameni, Anderut,
Bitrapar, Kiltan, Kadamum, Prain.
Urinaria, L.
Hedufuri, Goifurfehendu Atoll, Gardiner !
A cosmopolitan tropical weed. Minikoi ; Laccadives, in
Kalpeni, Prain.
96
WILLIS AND GARDINER : BOTANY
*\iruri, L.
Malé, Trimen ; no specimen. Goifurfehendu ik-toll, Vei-
mandu in Kolumadulu, Gardiner !
An almost cosmopolitan tropical weed. Minikoi ; Lacca-
dives, in Anderut, Akati, Kadamum, Kiltan, Prain. Diego
Garcia, Hemsley.
(rotuildifolius, Klein ; in Kiltan, Laccadives, Prain.)
Distichus, Muell-Arg,; Gobili, M. ; Rata-nelli, S.
I. Didi, 82 (Prain) !
Cultivated in Tropical Asia and the Mascarene Is. and in
Minikoi.
Glochidioii, Forst.
? littorale, Forst. ; Emboo, M.
I. Didi, 36, a leafy fragment, appears to be this species.
Croton, L.
Tiglium, L. ; the Croton-oil plant.
Veimandu in Kolumadulu, Gardiner !
Cultivated in the tropics for its medicinal seeds.
(Claoxyloii Mercurialis, Thw. ; in Akati, Bitrapar, Kada-
mum, Kiltan, Laccadives, and in Minikoi, Prain.)
Manihot, Tourn.
UTILISSIMA, Pohl. ; Cassava.
Male, Gardiner ! Only seen in Male. — J. S. G. Trimen ;
no specimen.
Cultivated in all hot countries.
Acalypha, L.
pauiculata, Miq. ; Dagundi kandi, Meia-sagu, M.
I. Didi, 169 ! Hulule 66, by paths through dense shady
jungle, Gardiner ! A weed of Tropical Asia and Africa.
iiidica, L. ; Vaffufuli, M. ; Kuppameniya, S,
I. Didi, 161 ! Male, Christopher, 1888 ! Goifurfehendu
Atoll, Addu Atoll, a very common weed. Gardiner !
A weed of Tropical Asia and Africa. Diego Garcia, Hemsley ;
Minikoi, Prain ; Laccadives in Akati and Kadamum, Prain.
OP THE MALDIVE ISLANDS.
97-
fallax, Muell.‘Arg. ; Bissagu, Cave (Minikoi), M.
I. Didi, 154 ! Goifurfehendu Atoll, in jungle outside the
mangrove swamp, Veimanduin Kolumadula, Gardiner ! Also
in Addu and Suvadiva Atolls. — J. S. G.
A common weed of Tropical Asia. Laccadives in Anderut
and Akati, Prain ; Minikoi, Prain, Gardiner.
Trewia, L,
? nudiflora, L, ; Varukundu, M.
I. Didi, 103, leaves only, appears to be this species. It
occurs in Ceylon, India, and Malaya.
Ricinus, L,
COMMUNIS, L. ; Amanaka, M. ; Endaru, S. ; Amanak-
kam, Chittamanaka, T. ; Castor-oil.
I. Didi, 143 ! Male, Gardiner, Christopher, 1888 ! Goi-
furfehendu Atoll, Hulule 45, Veimandu in Kolumadulu,
Gardiner ! Cultivated all over the islands. The seeds are
beaten up, and the natives oil themselves with the oil.
Cultivated in all hot countries. Laccadives, in Kiltan,
Anderut, Kadamum, Ameni, Bitrapar, Prain ; Minikoi, Prain,
Gardiner.
Urticacece. ,
Ficus, L,
benghaleiisis, LJ Nika, M. ; Mahanuga, S. ; the
Banyan.
sp, ? Eili, M. ; a Banyan.
Two banyans are found in Hulule, and there are large
banyans in Kenurus and other islands of N. Mahlos, well
back from the beach. Also in other atolls. Probably one
of these is F. benghalensis, and perhaps it is intentionally
introduced, as apparently in Ameni, Kadamum, and Minikoi
(Prain). A banyan is found in Diego Garcia (Hemsley,
?.c.334).
“ The long roots of these are used as poles for building
purposes, and also, I was told, for masts. Some of the trees
are very large, and must be of great age.”— J. S, G.
98 WILLIS AND GARDINER : BOTANY
Rbligiosa, L, ; Boi, M. (in this case always used with
the sufi&x gas=tree) ; Bo, S. ; the Bo or Peepul.
I. Didi, 32 ! “ One tree only, of large size, near the palace
at Male. There is said to be a second tree in Fua Mulaku, but
I did not see it.” — J. S. G.
Another is stated by Christopher in his report of 1836 to
be on the uninhabited Kedera I, in Suvadiva Atoll.
retusa, L. 9 or tsiela, Roxh, ?
I. Didi, 70, native name Kudehi, may be one of these. F.
retusa, L, ; in Minikoi, Prain.
infectoria, Roxh. ; Dumbu, M.
I. Didi, 17 !
Ceylon, rare.
Artocarpus, Forst.
INCISA, L. f. ; Bambukea (Banke in Addu Atoll), M. ;
the Bread-fruit.
I. Didi, 120 ! Hedufuri and Turahdu in S. Mahlos,
Limbo Kandu in N. Mahlos, Goifurfehendu Atoll, Hulule,
Gardiner ! At least a few trees in nearly every inhabited
island of the group. The fruit is largely dried in segments,
and kept against famine.
Cultivated throughout the tropics. Laccadives, in Kiltan,
Ameni, Anderut, Akati, Prain ; Minikoi, Prain.
INTEGRIFOLIA, L. f. ; Sakkeyo, M, ; Kos, S. ; the Jak.
Trimen’s list, “ I saw a few seedlings in Male. I have
never heard of any coral islands on which this flourishes,
though attempts have been made to introduce it into nearly
every one in the Pacific.” — J. S. G.
Cultivated in the tropics. Anderut, Laccadives ? Prain.
Fleurya, Gaud,
interrupta, Gaud. ; Gakehebuli, M. ; Wal-kaham-
biliya, S.
I. Didi, 101 ! Veimandu in Kolumadulu, Gardiner !
A weed from Tropical Africa to Polynesia.
OP THE MALDIVE ISLANDS.
99
Pouzolzia, Gaud,
indica, Gaud, ; Giteyokoli, Gitakoli (Minikoi), M.
I. Didi, 95 ! Goifurfehendu Atoll, Veimandu in Koluma-
dnlu, Gardiner ! Not seen elsewhere except at Malé. —
J. S. G.
A common, weed, Ceylon to China. Laccadives, in Kalpeni,
Kiltan, Akati, Kadamum, Prain ; Minikoi, Prain.
(Casuarina equisetifolia, Forst; in Diego Garcia, Hemsley.)
MONOCOTYLEDONS,
(The entire absence of Orchids, both here and in the
Laccadives, may be noted.)
SCITAMINEÆ.
Musa, L,
Sapientum, L. ; Keo, Dongkeo, M. ; Kehel, S. ; the
Plantain (Banana).
Tiladummati Atoll is famous for its plantains. Those
of Limbo Kandu, in N. Mahlos, are of good quality and of
luxuriant growth. They are cultivated in most islands,
often in pits two or three feet deep. See notes on Suvadiva,
below.
Cultivated in all hot countries. Laccadives and Minikoi,
Prain.
BROMELIACEÆ.
Ananas, L,
SATIVUS, L. ; Anana, M., S. ; the Pineapple.
Cultivated in Furadu Island of Ari Atoll, but of poor
quality. — J. S. G.
AmarylUdaceœ,
Crinum, L.
asiaticum, L,
Goifurfehendu Atoll, Gardiner (Prain)! Everywhere in
the Archipelago, sparsely, especially near to swamps with no
mangroves. — J. S. G.
(14)
100
WILLIS AND GARDINER : BOTANY
A littoral species, common on the coasts of Ceylon and
India.
Pancratium, L,
zeylanicum, L. ; Wal-lunu, S.
Wiligili in Suvadiva, Gardiner ! In damp jungle near the
village of Wiligili ; seen also in Kondai and Gaddu in the
same atoll. — -J. S. G.
Common in grassy places in Tropical Asia. Minikoi, Prain.
sp. ?
Didi, 11, native name Gul-hajaru, and Didi, 113 (Nargis
M.), appear to be fragments of spp. of Pancratium.
(Agave vivipara, L. ; cultivated in Anderut and Kiltan,
Laccadives.)
Taccacece.
Tacca, Forst.
pinnatifida, Forst. ; Hitala, Hittala,* Heith-thala, M. ;
Garandi-kidaran, S. ; Arrowroot.
I. Didi, 57 ! Goifurfehendu Atoll, Hulule 15, Gardiner !
Common in the whole Archipelago. The roots are scraped
and daily washed and dried in the sun for a week. They are
also used for poultices with chillies and cocoanut. They
form one of the staple foods of the natives against famine.
It is nowhere cultivated, but grows wild in nearly every island
in the group of any size. The near proximity of coral rock
is almost essential for its proper growth.
A littoral species, cultivated in Tropical Asia and Polynesia.
Probably sea-borne to the Maldives. Cultivated in the
Laccadives, in Anderut, Chitlac, and Akati, Prain. Also in
Minikoi, Prain, Gardiner.
DIOSCOREACEÆ.
DiosCOREA, L. ; the Yam.*
PENTAPHYLL A, L. ; Kattella, M. ; Katawala, S. ; Yam.
* Trimen gives Hittala as the Maldivian name for the Dioscorea yams,
his informant evidently describing Tacca as a “ yam.”
OF THE MALDIVE ISLANDS.
101
1. Didi, 37 (Prain ) ! “I think this must be the common
yam, cultivated in large gardens, especially in Tiladummati,
whence it is exported all over the N. part of the group. In
Addu it forms the chief food of the people.” — J. S. G.
Cultivated and semi-wild in the Old World tropics.
Alata, L»; Bihi-kattala,* M. ; Yam.
I. Didi, 39 (Prain) ! “ Cultivated in house compounds all
over the group : trained over fences and up cocoanuts or
other trees.” — J. S. G.
Cultivated in Tropical India.
Fasciculata, Roxh. ; Kurukuru, M. ; Yam.
I. Didi, 140 (Prain) !
Cultivated in Lower Bengal.
Globosa, Roxh. ? Mativa, Mathe-wa, M.
Rubella, Roxh. ? Kattala, M.
Purpurea, Roxh. ? Bilek-kattala, M.
I. Didi, 158, 42, and 159, mere fragments, are possibly
these species, Dr. Prain suggests. Cultivated in Bengal. “ I
do not know these.” — J. S. G.
(Sativa, L. ; cultivated in Ameni, Laccadives.)
(bulMfera, L. ; in Minikoi, Prain.)
Liliacece.
Asparagus, L,
racemosus, Willd. ; Satavaru, M. ; Hatawariya, S. ;
Chattavari, T.
I. Didi, no number (Prain) ! “ I saw this or a closely allied
form in several house compounds in Malé.”— J. S. G. Ceylon
Tropical Africa to Australia. Almost certainly intentionally
introduced.
Gloriosa, L.
Superba,L. ; Wihelia-gundi, M. ; Niyangala, S. ; Ven-
tonti, S. ’
I. Didi, 156 ! Male, Trimen ! Hedufuri, Goifurfehendu
Atoll, Hulule 9, Gardiner ! On kitchen middens, all over
the northern half of the Archipelago.
* So called from the leaves resembling those of the betel.
102
WILLIS AND GARDINER : BOTANY
Tropical Africa to Cochin China, Ceylon, Andamans,
Nicobars, Anderiit (Laccadives). Often littoral, e.^., in
Ceylon. Probably sea-borne.
Gommelinacece,
Commeliua, L,
beng;lialeiisis, L, ; Diyamudoli, M. ; Diyameneriya, S.
I. Didi, 2 ! Male, Gardiner ! Malé only, near mosques. —
J. S. G.
A grass weed, Tropical Africa and Asia. Possibly inten-
tionally introduced.
Kurzii, Clarke ; Tilo lagundi, M.
Hedufuri, Hulule 41, Kaddu in Haddumati, Gardiner !
Common in shade by villages in all S. Mahlos, and in Suva-
diva and Addu Atolls.
A weed of Ceylon and S. India ; not common,
sp. ?
Didi, 40 and 171 (Dikkunfati, M.), and 24 (Vakkunfati, M.),
are probably species of Commelina.
(Aneilema ovalifolium, Hk,f.; in Minikoi, Prain) !
Cyanotis, Don.
cristata, Schultes /.
Kaddu in Haddumati, Gardiner !
A weed in grass, Mascarene Is., Ceylon, India, Malaya,
Minikoi, Prain ; Laccadives in Kadamum and Kiltan, Prain.
Rhæo, Hance.
discolor, Hance.; Re Kandolu, M.
Hulule 29, Gardiner ! “ In a patch of jungle in the middle
of the cocoanut area : seen nowhere else except in Addu ;
quite wild. The thick lower stem roots are boiled and rubbed
on to cure rheumatism and sciatica.” — J. S. G.
PalmcB.
Arbca, L.
Catechu, L. ; Femfora, M. ; Puwak, S. ; Areca Palm,
Betel Nut.
OF THE MALDIVE ISLANDS.
103
‘‘In the eastern islands of Miladumadnlu, Malé, scarce
in Suvadiva, luxuriant in Addu. The nut is called Fo or
Fuakou.”— J. S. G.
Cultivated in all tropical countries. Ameni, Anderut, Kil-
tan, Minikoi.
Cocos, L,
nucifera, L, ; Khari (Kairi in Addu) ; another variety
Bo, M.; Pol, S.; Tenga, T. ; the Cocoanut.
“ Cultivated and wild in practically all the islands other
than mere sandbanks. There are several kinds of the oil
nut (khari) in cultivation, distinguished by their length and
shape. They are noted for great size in Addu and Mulaku
and to a less extent in Suvadiva as compared with other
atolls. They are always exported in the shell, copra not
being made. Addu nuts fetch in Calcutta about Rs. 40,
Northern Atoll and Minikoi nuts about Rs. 25, Nicobar nuts
about Rs. 45 per thousand. Another variety, of which the
flesh near the attaching stem is quite sweet, is grown in a
few islands and called Bo (head). The small sweet yellow
nut (kurumba) is probably only a recent introduction from
Ceylon ; a few are grown in the villages for drinking only.
The palm, as in other tropical countries, supplies many of
the necessaries of life, food, fibre, thatch, &c. Its stems, &c.,
are largely used in boat-building.” — J. S. G.
It is impossible, with the knowledge at our disposal, to say
whether this species was introduced deliberately by man, or
by the sea. The latter is perhaps more probable, as the
islands (see below) seem to lie so much in the track of the
sea-borne species.
Pandanacece.
Pandanus, L. /.
odoratissimus, L, /.
sp.? (Hornei, Balf. f. ? tall tree, leaf 15 ft. long.)
sp.? (Leaf spineless, 2 ft. long.)
104
WILLIS AND GARDINER : BOTANY
There appear to be three species in the islands : a very
large green-stemmed Pandanus (? P. Hornei, Balf. /., of the
Seychelles, P. Leram, Jones, of the Nicobars), reaching a
height of over 50 ft., found in Hulule, Turadu, Goifur-
fehendu. Limbo Kandu, &c. (Gardiner), and known as
Karikeo (Keeva in Addu Atoll) ; a small-leaved many-rooted
species, Bokeo (Divihe keeva in Addu), in Turadu, &c. ; and
a third, Medu Mokaneo, in Turadu, &c., smaller than either
of the above and more rooted and spreading. Trimen men-
tions P. odoratissimus under the name Ma-kahikeyo, but no
specimen is extant. This species is found in Minikoi, and
in Bangaro, Kadamum, Ameni, Kiltan, and Kalpeni in the
Laccadives, Prain.
“ The leaves of the Karikeo are drawn through the flames
of a fire to kill them, and split horizontally into two halves,
the spines being carefully cut otf. They are then scraped
and dried in the sun, to be steeped subsequently several
times in salt water and re-dried. Thus prepared they are
used in Tiladummati for making the large mat sails (rie%
and in every part of the group, especially Addu, for coarse
mats (santi) and pillows.
“ The natives are very fond of chewing the segments of
the fruit of the P. odoratissimus, and the ground under' a
dropping fruit is a great resort for hermit crabs {Gœnohita),
which also climb the branches.” — -J. S. G.
ARACEÆ.
COLOCASIA, L.
ANTIQUORUM, Schott, ; Ala, M. ; Gahala, S.
Sparsely throughout the Archipelago, Gardiner. Cultiva-
ted and more or less wild. “ Not regularly cultivated, except
in Tiladummati and Addu, and not much in the former. In
Maradu, Addu Atoll, there are plantations regularly irrigated
from wells. A second kind known by its lighter coloured
flesh is grown in some islands in brackish water, ^.e., in
artificially or otherwise lowered ground below high tide
OF THE MALDIVE ISLANDS.
105
limitSi It is seldom used, being kept against famine. It
appeared to me to be identical with the papoi of the Pacific,
Cyrtosperma edule, SchotV'* — J. S. G.
Cultivated in all hot countries. Minikoi ; Laccadives, in
Ameni, Anderut, Kalpeni, Kiltan, Kadamum, Akati, Prain.
Alocasia, Schott.
INDIÖA, Schott. ; Ma ala, M. ; Rata-ala, S.
I. Didi, 110 !
Cultivated in Tropical Asia.
Raphidophora, Schott.
PERTUSA, Schott.
Kaddu in Haddumati, Gardiner. A form with perforated
leaves, like that cultivated in Colombo gardens, probably
intentionally introduced.
Acorus, L.
Calamus, L.; Huvago, M. ; Wadakaha, S.
I. Didi, 15 (Prain) !
Cultivated in Ceylon and other warm countries.
Gyperacece.
Pycreus, Beauv.
pumilus, Nees. (Cyperus hyalinus, Vahl.).
Goifurfehendu Atoll, Gardiner ! Dry places near the
shore.
A weed of Ceylon, S. India, Timor, Laccadives in Kada-
mum, Prain.
polystachyus, Beauv. (C. B. Clarke) ! Kuna, M.
Midu in Addu Atoll, Gardiner ! A regular grass of swamps
in Suvadiva and Addu Atolls. Used for the well known
fine Maldivian mats.
Cosmopolitan in warm countries. Not recorded for the
Laccadives ; Minikoi, Prain.
Cyperus, L.
(compressus, L.; in Kalpeni, Laccadives, Prain.)
(pachyrhizus, Nees; in Bitrapar, Bangaro, and Kada-
mum, Laccadives, Prain.)
106
WILLIS AND GARDINER : BOTANY
(arenarius, Retz,^ a common littoral plant in Ceylon
and India, is absent from the Laccadives and
Maldives.)
(stoloniferus, Retz.^ also a common littoral species of
Trop. Africa, Mascarene Is., and Asia, is missing
also.)
(ligularis, L. ; in Diego Garcia, Hemsley.)
sp. ? Gokkuladuru, M.
I. Didi, no number. Perhaps a sp. of Cyperus, or the
following.
Ufariscns, Valil,
Drcgeaiius, Kunth. (Cyperus duhius, RottK) ; Hui
ru, Kalandura gundi (Minikoi), M.
Hulule 42, Gardiner ! One of the first to appear on any
sandbank that forms, outside the Scævola.
A littoral and inland species, Ceylon, India, Malaya,
S. Africa. Evidently sea-borne, as indicated in the note
above, and further described below. Minikoi, Prain,
Gardiner.
albescens, Gaud. ? (Cyperus pennatus, Lam.).
Goifurfehendu Atoll, Gardiner. Almost certainly this
species, but a very reduced specimen. Near the shore, on
dry sandy places.
A littoral and inland species, Mascarene Is. to N. Australia,
Minikoi, and Laccadives, in Andern t, Kalpeni, Akati, Prain.
(Ryllinga brevifolia, Rotth. ; in Minikoi, Prain.)
(R. monocephala, Rotth. ; in Diego Garcia, Hemsley.)
Fimbristylis, Vahl.
spathacea, Roth.
Trimen’s list ; no specimen. Goifurfehendu Atoll, Vei-
mandu in Kolumadulu, Gardiner ! Dry places near the
shore. Sometimes used for mat-making.
A littoral species, common in Ceylon, India, Malaya, &c.
(diphylla, Vahl.; in Kalpeni, Laccadives, Prain.)
(glomerata, Nees ; in Diego Garcia, Hemsley.)
OP THE MALDIVE ISLANDS.
107
Cladium, P. Br.
jamaicense, Crantz {Mariscus, R. Br.).
Addu Atoll, Gardiner. (Prain ! “ As nearly as possible
typical,” C. B. Clarke ! ) Grows to a height of 4-5 ft., almost
completely covering a large swamp in Midu island (water
quite fresh). — J. S. G.
A cosmopolitan species, temperate and tropical. Not
recorded for India or Ceylon, except from high elevations
in Kashmir ; Mauritius.
[Remirea maritima, Atihl., a cosmopolitan tropical littoral
species, common in Ceylon, is absent from the Maldives and
Laccadives, so far as we know.]
Gramineœ.
Paspalum, L.
sang'uinale, Lamk.
Goifurfehendu Atoll, Gardiner ! Common in Suvadiva.
Cosmopolitan in warm countries, Minikoi, Akati (Lacca-
dives), Prain ; Diego Garcia, Hemsley.
Panicum, L.
Miliaceum, L.; Bai, M.; Wal-meneri, S. ; Millet.
Trimen’s list ; no specimens. See Bell, lx. 84. Culti-
vated. “ Formerly extensively cultivated throughout the
whole group : only saw it in Kenurus (Mahlos) and Mafaro
(Miladumadulu).” — J. S. G.
(trigonum, Retz.; Hue, in Minikoi, Gardiner.)
Spiiiifex, L.
squarrosus, L.
Common in N. Mahlos, Goidu (Horsburgh), &c. — J. S. G.
A littoral species, on sandy shores, Ceylon, India, Burmah
to China, Minikoi, Prain ; Laccadives, in Bitrapar, Prain.
Sea-borne.
Oplismeiius, Beauv.
' (Burmanni, Beauv.; in Akati, Kadamum, Kiltan,
Minikoi, Prain«)
compositus, Beauv. ; Ona hui, M.
(15)
108
WILLIS AND GARDINER : BOTANY
Hedufuri, Hulule 37, Veimandu inKolumadulu, Gardiner !
A very common wet jungle grass, especially in Addu and
Suvadiva. — J. S. G.
An almost cosmopolitan tropical grass. Laccadives, in
Ameni and Akati, Minikoi, Prain.
Setaria, Beauv,
ITALICA, Beauv,; Ura, M. ; Italian Millet.
Trimen’s list ; no specimens. “Not so common at any
time as Panicum miliaceum. I was given a little at Raska-
teen, N. Mahlos, where some was growing half wild.” —
J. S. G. Cultivated in Ceylon and other warm countries.
(verticillata, Beauv. ; in Ameni, Bangaro, Kadamum,
Kiltan, Prain.)
Stenotaphrum, Trin.
complanatum, Schrank.
Veimandu in Kolumadulu, Gardiner !
Tropical Africa, Mascarene Is. to N. Australia. Diego
Garcia, Hemsley. Perhaps sea-borne, like other grasses.
Tlmarea, Pers.
sarmentosa, Pers.
Hedufuri, Goifurfehendu Atoll, Gardiner ! Common on
sandy sunny beaches throughout the Archipelago ; character-
istic of the same area in Goidu as Ipomcea biloba. — J. S. G.
A littoral sea-borne species, Mascarene Is., Ceylon, Nico-
bars, Andamans, Malay Is., Polynesia. Not in India.
Minikoi, Laccadives in Bangaro and Kadamum, Prain.
(Oryza sativa, L. ; Rice ; Handu, M. ; appears to be
occasionally cultivated in some islands, but the bulk of the
supply is imported from India. “ The natives have tried to
introduce this many times, but it will not grow and set
seed.” — J. S. G. Cultivated in Anderut, Laccadives.)
Zoysia, Willd.
puugens, Willd.
Trimen ; no specimens. A common littoral grass in Trop.
Asia, Australia, and Mauritius.
OF THE MALDIVE ISLANDS.
109
Zea, Z.
Mays, L, ; Donera, Donala, M. ; the Maize.
I. Didi, 35 ? Cultivated in Landu, Gardiner.
Cultivated in all warm countries. Not in Minikoi or the
Laccadives.
Saccharum, L.
OFFICINARUM, L.; Udandi, M. ; Sugar-cane.
Trimen ; no specimens. “ A little grown in house en-
closures in nearly all the larger islands of the group ; may
almost be said to be cultivated on a large scale in Landu and
Mahugudu (both in Miladumadulu).”^ — J. S. G.
Cultivated in all warm countries. Minikoi, Prain.
Ischæmiim, L.
ciliare, Retz,; Bileh-hui, M. ; Rattana, S.
I. Didi, 126 (Prain) !
A common grass, Ceylon to Australia and China. Lac-
cadives, in Kalpeni, Akati, Bitrapar, Kadamum, Kiltan,
Prain.
miiticum, L,
Hedufuri ; very common along the S.E. line of islands of
S. Mahlos on broad roads which have gone back to jungle ;
very sparingly elsewhere ; Goifurfehendu Atoll, Gardiner ! .
Common from S. India to Australia. Laccadives in
Kalpeni ; Minikoi, Prain.
Apluda^ L.
varia. Hack., var. aristata, L. ; Ona hui (Minikoi), M.
Hedufuri, Goifurfehendu Atoll, Hulule 8, Veimandu in
Kolumadulu, Gardiner ! Covers a large area in Hedufuri
in a damp low area in Beembi land. The commonest grass
by far in Hulule. Common in the whole group. — J. S. G.
A common grass from Ceylon to Polynesia. Minikoi,
Laccadives in Kadamum, Prain.
[Cynodon Dactyloii, Pers.; in Kelpeni, Laccadives.]
110
WILLIS AND GARDINER : BOTANY
Andropogon, L.
Sorghum, BroL ; Zoowari, M. ; the Juar or Jowari of
India.
Formerly cultivated in the Maldives. — J. S. G.
Squarrosus, /. ; Lancimo, Lunsimoo, M. ; Sae-
wandara, S. ; the Khus-khus.
I. Didi, 29 ! Fainu in N. MahloSj Gardiner (Prain) !
Extensively cultivated in Suvadiva. The scented roots used
as ornaments for the women’s hair.
Cultivated, andwildin the Old World tropics. Laccadives,
in Kiltan, Prain.
[contortus, L, ; in Kiltan and Kadamum, Laccadives,
Prain.]
N ARDUS, L. ; Kasinji, M. ; Mana, S. ; Citronella Grass.
I. Didi, 79 (Prain) !
Wild and in cultivated forms. Old World tropics.
Eleusine, Gaertn.
[indica, Gaertn.; in Minikoi, Prain.]
CORACANA, Gaertn.; Bimbi, Beembi, M. ; Kurakkan,
S.
Formerly the most largely cultivated grain in the Archi-
pelago, but now only a little grown here and there. The
area to be planted is cut down and all burnt off by March.
The surface is then smoothed over, and the grain is scat-
tered by hand after the first rain of the S.W. monsoon.
The crop is ready to cut in four months. The heads only
are taken, the straw being left and reburnt.” — ^J. S. G.
Cultivated in Ceylon, India, Egypt, Japan, &c. Minikoi,
Laccadives, in Ameni, Anderut, Kadamum, Kalpeni, Prain.
Æ^yptiaca, Desf.
Trimen ; no specimens. Goifurfehendu Atoll, Gardiner !
Common in cleared Bimbi land, especially in Miladum-
madulu. — J. S. G.
OF THE MALDIVE ISLANDS.
Ill
A weed, common in the Old World tropics, and natu-
ralized in America. Laccadives, in Ameni and Kadamum,
Prain.
Era^rostis, Host.
tenella, R. & S., var. pliimosa, Stapf. ; Tubuli hui,
Sannipoo (Minikoi), M.
I. Didi, 141 ! Goifurfehendu Atoll, Hulule 10, very com-
mon, Gardiner ! One of the first grasses to reach a sand-
bank, followed by Scævola.
Common in Tropical Asia and Africa. Sea-borne. Mini-
koi, Prain, Gardiner.
Laccadives, in Kalpeni, Akati, Kadamum, Kiltan, Prain.
(Lepturus repeilS, R. Br.; in Minikoi, and Laccadives in
Bangaro and Bitrapar, Prain. Littoral, a rare grass in
Ceylon, Malay and Polynesian Is.)
Bambusa, Schrei).
Arundinacea, Willd., or Vulgaris, Sehr ad., or
both ? or more ?
Malé and other islands. A plain yellow bamboo, ap-
parently vulgaris, is grown in Landu. A dwarf kind is
more common, growing in many house compounds. — J.S.G.
GYMNOSÏERMÆ.
Cycadaceœ.
Cycas, L.
circinalis, L.
Island near Malé, F. Lewis.
, CRYPTOGAMÆ.
Filices.
(Asplénium æquabile, Baker.
longissimum, Blume. -k
falcatum, Lam. ; all in Diego Garcia, Hemsley.)
Aephrolepis, Schott.
exaltata, Schott. ; Kees fila, Makunu hungal, M.
112
WILLIS AND GARDINER : BOTANY
I. Didi, 168 ! Hedufuri, Hulule 30, near the roots of the
cocoanutsin damp sandy areas, Veimandu in Kolumadulu,
Kaddu in Haddumati, Gardiner (Prain) ! Also in Suva-
diva.— J. S. G.
Cosmopolitan ? in the tropics.
(tuberosa, Presl, ; in Minikoi, Gardiner.)
(cordifolia, Presl. ; in Anderut, Laccadives, Prain.)
Tliamnopteris, Presl.
Mdus, L.; the Bird’s Nest Fern.
Veimandu in Kolumadulu, Kaddu in Haddumati, Gar-
diner ! Also in Suvadiva. — J. S. G.
Common in the tropics in very damp localities. Diego
Garcia, Hemsley.
Nephrodium, Rich.
molle, Desv.
Kaddu in Haddumati, Gardiner !
Cosmopolitan in the tropics. Laccadives in Anderut,
Prain.
(unitum, R. Br. ; in Diego Garcia, Hemsley.)
(Pteris marginata, Bor y ; in Diego Garcia, Hemsley.)
Lycopodinece.
Psilotum, Sw.
triquetrum, Sw. ; Prumo (Minikoi), M.
Goifurfehendu Atoll, Kaddu in Haddumati, Gardiner.
Cosmopolitan in the tropics. Minikoi, Gardiner, Prain ;
Diego Garcia, Hemsley.
“ In Goidu on the bases of cocoanuts close to the mangrove
swamp. In every island of Suvadiva. Grows chiefly on
cocoanuts, on the top of the basal mass oî roots or in holes
in the trunk.” — J. S. G.
Finally there are three utterly indeterminable fragments : —
I. Didi, 38, Dakada, M.
67, Bambara, M.
157, Makumburuvani, M.
OF THE MALDIVE ISLANDS.
113
IIL— NOTES ON
THE VEGETATION OF THE VARIOUS ATOLLS.
(By J. S. Gardiner.)
Hulule Island.^
This is the S.E. island of Male Atoll. On the E. and S.
the reef is one-third to half a mile wide, on the W. about
100-150 yards, on the N. the island lies along the line of the
reef. The soil is sand, very fine to the west, but to the east
foraminiferal and coarse, with a little pumice. There
are coral blocks in the sand along the cocoanut belt south
of the village. The rainfall is probably about 110 inches ;
certainly considerably less than in islands lying further
west and south.
Going to the land for a moment it is interesting to notice
the distribution of some of the plants and trees. In the
first place, the cocoanuts form a belt from the N.W. point
of the island down to the village, and then along and across
the island to the undermined beach. East of this — that
is, a belt along the N.E. shore of the island — only small
bushes occur, and no trees of any size. On the west side
of the village is a belt of low scrub 30-40 yards broad, and
south of the village is a dense jungle of Bambukeo (bread-
fruit), Funa (Calophyllum Inophyllum), Karikeo (Pan-
danus Leram ?), Kandu (Hernandia peltata), two or three
candle-nut trees (Barringtonia speciosa), with an odd
Hibiscus tiliaceus here and there. The south point of the
island has no large timber, and the whole west and south
shores are fringed with Kuredi (Pemphis acidula), with an
occasional Mabori (Tournefortia argentea) here and there.
By the village is a zære (holy grave), and behind this is a
small untimbered patch, really an old kitchen midden with
I spent nearly six weeks on this island during January and February,
1900. One boy was employed solely in collecting and looking after the
flora, always accompanying me. Although out daily I failed to add any
plants in the last three weeks of our stay.
114
WILLIS AND GARDINER : BOTANY
a large clump of jessamine (Jasminum grandiflorum) in
the centre, the rest consisting principally of a yellow
composite (Wedelia biflora), low shrubs, and climbing beans.
The whole west and south sides of the island, where the
Pemphis forms a single line of timber on the top of the
sandy beach, are slowly washing away. The same action is
more rapid to the south-east, where fallen cocoanut palms
strew the beach. Further north the beach is stationary
or slightly growing out by the washing up of sand, which
reaches its maximum in a broad blunt sandy point to the
N.E. Pemphis gradually gives place to Scævola, with a
clump of Tournefortia here and there, until these latter are
alone found. At the point these shrubs have not been
able to keep pace with the beach, and there is outside the
line of advancing Scævola a low growth of Mariscus Drege-
anus, Launæa pinnatifida, Eragrostis plumosa, with very
numerous seedlings of Tournefortia, and a few plants of
Aerua lanata. Behind, the Scævola line is very marked,
merging in 30-40 yards into a low bash — continuous up to
the cocoanut belt — formed almost entirely of Ochrosia
borbonica and Morinda citrifolia, weighed down in many
places with Gassy tha filiformis.
The flora of Hulule, which lies very close to Malé Island
itself, and thus in the track of trade, contains, as might be
expected, many introduced weeds of cultivation. Apart
from the strictly cultivated forms, it consists of 58 species,
of which perhaps about 22 are man’s weeds and about 24
littoral species, sea-introduced.
Goifurfehendu (Horsburgh) Atoll."'
The atoB consists of five islands : Goidu, Fehendu,Furudu,
Inafuri, and Doru Kandu, the plants of which decrease in
variety in the above order. Doru Kandu lies to the south,
close to the outer part of the reef, and is all rocky ; it is
* Our visit was in November, 1899, The plants were carefully col-
lected by myself alone. So far as the small herbaceous species are
concerned, the collection is about complete.
OF THE MALDIVE ISLANDS.
115
destitute of all vegetation, except a few old Pemphis trees
half growing in the water at the east end. All the other
islands are formed entirely of sand, with the exception of
Goidu, which has a rocky belt 30-40 yards broad at its
extreme eastern end. This is largely fringed near the sea
with Pemphis, while cocoanuts, small Pandani, Allophylus
Cobbe, Hernandia peltata, Acalypha fallax, and Premna
integrifolia form a dense jungle behind. The rocky and
sandy areas join in a small mangrove swamp, west of
which is a dense jungle of bread-fruit, cocoanuts, and
large Pandani, with a few large banyans and Terminalia
Catappa. The west shore of Goidu forms a slight bay, and is
fringed with Pemphis acidula and Tournefortia argentea N.
and S., and in the centre Scævola Koenigii and Suriana
maritima, with a small growing area outside covered with
dense Spinifex squarrosus. Contrasted with Goidu the
other islands are noticeably poor in low herbaceous plants.
Inafuri is merely a sandbank covered with low trees and
bushes, Spinifex and Launæa being the only herbaceous
plants observed.
Of food plants the cocoanut is alone of any importance,
covering the greater parts of Goidu, Fehendu, and Furudu ;
plantains, yams, and grain of any kind do not grow, and the
papaw only reaches a very small size. The flora of this atoll,
collected on Goidu and Furudu, with a few plants from
other spots, numbers, apart from the strictly cultivated
forms, 67 species.
Mahlosmadulu Atoll.
This is divided into North and South Mahlos Atolls.
Plants were collected from several islands in each.
Turadu,"^ in South Mahlos, is a small thickly inhabited
islet with no indigenous foliage or jungle, the greater part
planted with kurumba (yellow edible cocoanut), bread-fruit,
* Only a few plants, picked as not found or not common in Goifur-
fekendu.
(16)
116
WILLIS AND GARDINER : BOT AN
papaw (falo and veofaio), Barringtonia speciosa, and various
kinds of Pandanus (Karikeo, Bokeo, Medu Mokaneo), with
various herbs used in curries. The whole of the south of
the island is being very rapidly washed away by the sea,
and there is a long point extending outwards, covered with
Pemphis acidula, the stems of which are washed for two or
three feet at high tide, and are very slowly being killed ;
considerable changes took place here two years and five
months before I visited the island, since when the natives
say that they cannot see any further change.
Hedufuri, in S. Mahlos, is likewise a small thickly inha-
bited very dry sandy islet. The village is situated on the
east side, and behind it is a small clump of cocoanuts, bread-
fruit, &c. The rest of the island is very dry, sandy, covered
with a low growth of Scævola, Ochrosia, Tournefortia, and
small Pandanus, with grassy patches here and there.
Morinda citrifolia is very abundant, being to some extent
planted by the people for the root, from which a permanent
red dye for cloth is obtained by mixing a decoction with
lime.'*^ The flora of this island numbers 46 species other
than cultivated forms.
The islets along the south side of Mahlos Atoll may be
divided by the vegetation at a glance into sandy and stony
islets, or, if mixed, the two parts can be separated.
The sandy islets have, as a rule, a few semi-dwarfed cocoa-
nuts and bread-fruit trees in the centre, surrounded by such
low bushes as were mentioned underHeduf uri,with Scævola
beaches.
The rocky islets, on the other hand, are covered with dense
tall jungle of cocoanuts, large Pandani, Calophyllum, Bar-
ringtonia, with often a clump of banyans, and have Pemphis
beaches.
* Probably some of the other plants were introduced as dyes and have
now run wild. The whole population is engaged in cloth-making. The
black thread is now imported as such, but was formerly dyed, the colour
being obtained from the sap of the plantain. The collection from this
island is fairly complete.
OF THE MALDIVE ISLANDS.
117
A marked increase in the size and luxuriance of the vege-
tation is noticeable in passing from the west to the east of
Mahlos Atoll.
In North Mahlos the islands of the east side are formed
for the most part of rock toward the seaward face with
coarse sand inside. Compared with S. Mahlos, there did
not appear to me to be nearly the variety in herbaceous
plants, while woody plants are the same, the trees generally
much taller and often of quite remarkable size. Kenurus
in particular is covered in places with large banyans and
Barringtonias. They everywhere keep well back from the
beach, and avoid the sandy area, growing most luxuriantly
at the line of junction of rock and sand.
The western and inner islands of N. Mahlos resemble the
sandy islands of S. Mahlos for the most part ; the majority
are sandbanks forming or being washed away. Of these, I
visited over thirty, and there seemed to me to be a definite
order in which the plants came to them. All are at some
time the resort at high tide of large flocks of terns, herons,
and sandbirds, the common crow appearing only when they
become inhabited. Herbaceous plants first appear, most
commonly Spinifex squarrosus or Launæa pinnatifida, to be
followed by Mariscus Dregeanus, Eragrostis plumosa, and
perhaps Aerua lanata. Very quickly seedlings of Tourne-
fortia argentea appear, and often these grow into large
bushes before the Scævola arrives, and indeed this plant
often remains for a long time absent, but as soon as it settles
it at once spreads with great rapidity, only a relatively old
Tournefortia tree being left here and there. On the first
bushes Gassy tha filiformis is sure to be found. Ochrosia
borbonica, Terminalia Catappa, and small Pandani soon
appear, and in large sandbanks the central part is often
covered with Guettarda speciosa and Hernandia peltata.
Cocoanuts do not, as a general rule, grow well on these
islets, and the largest have seldom more than a few straggly
118
WILLIS AND GARDINER : BOTANY
trees in the centre. If the islets are washing away at all,
Pemphis acidula is sure to be present above the beach, but
if they are growing is very seldom found.
There are very few rocky islets in the whole group which
did not exist when the charts were made in 1836, but Hura,
to the west of the reef marked Wah-pure, was then devoid
of foliage. It has now cocoanut trees said to have been
planted about twenty years ago. before which it was all grass.
Now Pemphis acidula and Tournefortia argentea occur about
the beach abundantly to the east, while the west, where the
island is still growing, is covered with Apluda aristata,
Aerua lanata, with Canavalia lineata, Oldenlandia biflora, and
Cassytha flliformis creeping over the rocks in every direction.
The absence of Pandani is remarkable.
Limbo Kandu, in N. Mahlos, is remarkable as showing
the richest vegetation we saw north of Male. It is a round
sandy island with a ridge of wind-blown sand all round,
perhaps 12 feet above high tide level, below which there is
little change taking place in the beach. Round the coast is
a line of the ordinary sand bushes, 80-120 yards broad, in
which the ridge above slopes to the central flat, which is 1-2
feet above high tide. This part is for the most part covered
with large trees, cocoanut, banyan, bread-fruit, Barringtonia,
with a number of very large limes and Pandani. Plantains
of very good quality grow luxuriantly, and there are some
yams (Ipomœa Batatas), pomegranates, a little ala (Colocasia
antiquorum), and best of all, chillies. A noticeable point
in the central clump is the extreme luxuriance of a certain
creeper which has completely covered over some large
Pandani and hangs from the Nika (banyan) in heavy festoons.
The tropical appearance was only perhaps equalled in
Addu Atoll, but the creeper was not seen elsewhere. It
should be mentioned that the island was formerly inhabited,
and the creeper may have been brought from Ceylon or
India with the plantains, which are frequently brought over
OF THE MALDIVE ISLANDS.
119
fresh for planting. Contrasting Mahlos with other atolls,
the comparative absence of yams, plantains (with the sole
exception of Limbo Kandu), ala (Colocasia), funa (Calo-
phyllum Inophyllum), papaws, and indeed of all food plants
is very noticeable. Hitala (arrowroot, Tacca pinnatifida)
occurs in nearly every large island, but in no luxuriance ;
here only in the group is it regularly collected and eaten,
the same remark applying to the mangrove also where it
occurs. The cocoanuts are generally of small size, and the
number per tree is considered by the natives very small.
Generally the soil would not appear to be nearly as rich as
in the eastern line of atolls.
S. Mahlos has been for many years closely connected
with Male, from which the inhabitants have obtained rice ;
there is only evidence of grain having been grown in
Doomfang in recent (about twenty-five) years. In N. Mahlos
grain (Eleusine Coracana and Panicum miliaceum) seems to
have been grown in nearly all the islands within the last 8~10
years : in 1899, however, it was only grown in Kenurus,
planted with the first rains of the little monsoon in April,
and reaped in August. To plant it, the grass and shrubs
being burnt off, the whole was surrounded by a cadjan fence
1-2 feet high to keep off the rats, the surface smoothed by a
rake and the grain strewn on by hand. None of the
western islands of N. Mahlos are inhabited, but a few inner
ones have villages of manufacturing or fishing castes.
Ari and filEandu AtoEls.
The descriptions of North Mahlos apply equally well to
these atolls. The Island of Furadu produces small pineapples
of poor quality and flavour. We coasted Ari close to the
eastern shore for 15 miles, and steamed through N. and S.
Nilandu, visiting, however, only Rimbuduin S. Nilandu and
Mahikaddu in Ari, which in the general appearance of their
foliage, as well as in the plants, exactly resembled islands in
the same position in N. Mahlos.
120
WILLIS AND GARDINER : BOTANY
Fadiffolu, N. and S. IVEale, Felidu, and
Mulaku Atolls.
Hulule in Male is described separately ; it is probably
richer in variety of indigenous plants than any of the other
islands, on account of its large size, position, small popula-
tion of low caste, and little cultivation.
The atolls mentioned lie along the eastern side of the
Archipelago. The islands of their western side are not
inhabited, except in Fadiffolu, and are generally bare and
sandy, with low trees and bushes only ; cocoanuts of small
size are scattered over them, poor in quantity and in quality.
The eastern islands are generally densely covered in cocoa-
nuts grown for the nut and sugar, but plantains, yams, and
vegetables of all sorts do not flourish, and are not found
anywhere in luxuriance. All appear from the sea to be flat-
topped high cocoanut islands fringed with Pemphis acidula
generally ; banyans, Calophyllum,&c.,very scarce ; the Areca
palm not found.
Kolumadulu and Haddumati Atolls.
None of the islands of these atolls were visited by me,
except Buruni in Kolumadulu ; a complete set of plants was
collected on Yeimandu in Kolumadulu, and a few from other
places in that atoll. A collection was also made on Kaddu
in Haddumati. The eastern islands of both atolls are
densely covered with cocoanuts, but have a good deal of
high timber as well. Buruni does not differ greatly from
the generality of Mahlos islands, but struck me as being
especially rich in small herbaceous plants and grasses ; all
seemed familiar, however, except a ‘‘ shamrock,” Desmodium
triflorum, which I saw nowhere else. The Yeimandu flora,
apart from cultivated forms, numbers flfty-five species.
Suvadiva Atoll.
As already mentioned, this atoll lies much to the south,
separated by a wide channel from Haddumati to the north,
and from Addu to the south. The eastern islands of this
OP THE MALDIVE ISLANDS.
121
atoll are nearly all densely covered with large cocoanut
trees, as are also some of the western islands, but many of
the latter have only low trees and shrubs. The inner islands
are, as a rule, covered with low trees and bushes, Heebahdu
to the north with very large Barringtonia speciosa ; cocoa-
nuts if present on these islands are of small size and
scattered. A few plantains and food plants are found in
every island, but they do not grow at all luxuriantly; the
former are now planted in pits in the ground dug to below
high tide level, and in Kondai I saw a man put in a basket
of pumice from the beach ; this appeared on inquiry to be a
regular custom of the island. A little grain is stated to have
formerly been cultivated, but the people seem to have been
for all time famous as daring navigators, and to have lived
for the most part on rice obtained from India in exchange
for cocoanuts, mats, and fish. Many of the islands of the
outer reef hâve swamps overgrown completely or fringed
with Kuna grass (Pycreus), used for their great industry of
mat-making.
Wiligili has one such swamp now almost dry, covered
with kuna, with Leucas biflora growing on the roots of the
trees and on any drier spots. A clump of Nephrolepis exal-
tata and Thamnopteris Nidus also occur here. Round this
the jungle is very thick and dense, the larger trees -consist-
ing for the most part of Cordia subcordata and Terminalia
Catappa, the undergrowth being composed almost entirely
of Ardisia humilis, with here and there a bush of Rhizo-
phora mucronata and Lumnitzera racemosa ; in certain
places one sees a bush of Hibiscus tiliaceus and occasionally
a giant cocoanut or Pandanus. The paths are fringed by
Oplismenus compositus, and near the village Pancratium
zeylanicum is very common.
On sandy islets the most striking feature compared to
other atolls is the dominance of the Pandani to the detri-
ment of Scævola and Tournef ortia ; Pemphis, too, is not
abundant.
122
WILLIS AND GARDINER : BOTANY
Collections of plants were made in Wiligili, where there
is a large village, and in Gaddu. A kind of small Pampas
grass called Lancimo (Andropogon squarrosus, the Khus-
khus) is grown extensively throughout the whole atoll. Its
fibrous root has a sweet somewhat lemon-like smell, and is
used as ornaments for the women’s hair. A kind of thyme,
too (tulamu), is grown for the same purpose.
Addu Atoll.
This little atoll, situated 35 miles south of the Equator, is
perhaps the most perfect of all I saw in the Maldives, having
only two narrow passages to the north and two broader ones
to the south. The encircling reef appeared from the ship to
be covered with land, which but for the passages seemed to
be continuous. It was really, however, broken up into
numerous islands by channels which often could not be
seen on account of smaller islets standing on the outer part
of the reef behind. The islands of the east side each consist
of sand and rock areas respectively against the lagoon and
open sea ; these areas are separated by a series of kuli or
fresh water lakes in the centre of the islands. On the west
side the sea has everywhere broken into these lakes, and
two series of islands result, rocky patches against the sea
and larger sandy islets by the lagoon. The difference
between the two sides is probably due to the heavier weather
which comes up from the west. The atoll is almost beyond
the reach of the Indian monsoons, the north-east monsoon
bringing only a few gales from the north in January and
February, but in the greater part of the year the westerly
trades prevail. The rainfall, too, being probably about
150 inches, is heavier than in any other part of the Maldives,
and perhaps this is the cause of the exceptionally rich
vegetation which characterizes the atoll.
North-east of the atoll is a large island with two densely
inhabited villages, Midu to the north and Huludu to the
south, having between them a population of about 3,000
OF THE MALDIVE ISLANDS. 123
people. Midu is entirely an agricultural village, but Huludu
is inhabited by high caste people and sailors who trade up
and down the group ; it usually also sends two or three ships
a year to the Hooghly, and has traditions of an Arabian trade
in times past. The village itself is full of fine houses with
large compounds, in which one may recognize almost as
great a variety of spices and other cultivated plants as at
Male. Every available inch of the island is cultivated,
except a narrow debatable strip between the two villages.
The people for the most part live on fish and on the vege-
tables they grow ; no grain was, so far as I could ascertain,
ever grown, and rice is not much imported. Cocoanuts are
planted right up to the beach on every side, and are of
greater size and luxuriance than any I have seen in Ceylon
or the Pacific. They also bear heavier crops both in number
and size ; 60 trees are supposed to yield 1,000 nuts worth
Rs. 45 in Calcutta, and about 4,750 nuts give a ton of copra.
The kurumba or yellow form is not grown, but a green form
with sweet husk is regularly cultivated for the crews of the
vessels. The centre of the island has a great fresh water
marsh overgrown with Cladium jamaicense, with Pycreus
polystachyus near the edges. The north part of this has
been reclaimed and turned into a great yam and sweet
potato flat, the latter dug four times a year. Between it and
the marsh long beds have been planted with Colocasia anti-
quorum, some of which is almost growing in the water.
Near the village are extensive plantations of plantains, of
which the people distinguish five varieties, and areca palms
grow everywhere in competition with the cocoanut. Each
house has its papaws and bread-fruits, and pumpkins, water
melons, chillies, and betel grow almost semi-wild in their
compounds. The debatable ground is covered with Hibiscus
tiliaceus struggling with Hernandia peltata, both of which
almost appear as if planted. The roads are everywhere well
kept ; they are edged for the most part with grass, flowering
dicotyledonous weeds being remarkably scarce and not as
(17)
124
WILLIS AND GARDINER : BOTANY
in the northern atolls struggling with and maintaining their
position against the grass.
The islands to the south of Huludu have the usual trees
and shrubs characteristic of rocky and sandy areas. Some
are planted with cocoanuts. Most are very barren, the
arrowroot (Tacca) being almost the only herbaceous plant
found besides the grasses. Convolvulus and other creeping
and climbing plants are practically non-existent on the shores
of either side. Scævola and Tournefortia do not dominate
any part. Pemphis grows against the sea, and has been
planted in lines across the kuli. For the rest it is an open
jungle of small Pandani and the usual shrubs, with the
addition of Gyrocarpus .Jacquini, Morinda citrifolia, and
Thespesia populnea. The absence of mangroves from so
suitable an area is most remarkable, it seems. to me.
Maradu, one of the central islands of the west side, has
perhaps an even more luxuriant vegetation than either Midu
or Huludu. It is more dank and moist, pools of water
standing everywhere. The village is shaded with immense
bread-fruit trees, and every house has its trellis work of
cucumbers, gourds, and pumpkins ; water melons grow
everywhere. Indeed, practically every vegetable which is
at all extensively grown in the group flourishes here, the
brinjals being as good as those of Ceylon. Behind the
village is a dense tall jungle of cocoanuts, arecas, Calo-
phyllum, Terminalia Catappa, Cordia subcordata struggling
with an undergrowth of limes, Zizyphus Jujuba, and many
other wild and semi-wild trees. The soursop, pomegranate,
Sonneratia, and betel grow without being cultivated, and
there are even said to be a few orange trees. Coarse grass
covers the untilled ground everywhere, save in the densest
shade. Herbaceous weeds are rare ; almost the only flowers
are the Rhœo discolor and the Pancratium zeylanicum.
There was one single immense Barringtonia speciosa, but
I do not remember to have seen a single banyan tree in Addu,
and I think it is not found.
OF THE MALDIVB ISLANDS.
1*25
Miladumadulu Afoli.
Only two islands on the west side of this atoll are at
present inhabited, while most of the outer western line of
islands have people on them or are regularly visited. The
western islands are generally covered with low trees and
shrubs, while the eastern islands have large cocoanuts and a
few banyans. The greater part of the eastern islands has at
times been cultivated, but very many have large lakes
(kuli, M. ; kulam, T.) in the centre, surrounded by a thin
fringe of mangroves, of which the natives distinguish three
kinds according to their edible properties, Bruguiera caryo-
phylloides being the most numerous.
Formerly these islands were nearly self-contained, all
growing grain, yams, sweet potatoes, plantains, pumpkins,
and Colocasia ; arrowroot (Tacca) is very plentiful. The
people seem to have been very industrious, and all the
twenty -eight islands we visited seem to have been at some
time cultivated. Compared with Mahlos Atoll, many of the
islands exhibit almost as luxuriant a vegetation as Limbo
Kandu, and surpass all the other islands of that atoll. Trees
and shrubs are nearly the same, but Pandani, bread-fruit, and
Galophyllum Inophyllum are dominant ; herbaceous plants
exhibit little variety in any one island. By the mosques
sweet-smelling plants are much grown, poor varieties of the
rose being dominant.
The stony eastern area of all these islands is generally
covered with cocoanuts, which prefer such a situation. The
intermediate vegetation consists of Pandani, Galophyllum
Inophyllum, Hernandia peltata, and Hibiscus tiliaceus, all
useful economic products.
Mahagudu and Landu are particularly noted for their
great variety of economic plants ; in the latter, in addition to
the above, the areca palm, betel, sugarcane, maize, sour-
sop, water melons, limes, lemons, chillies, Sonneratia acida,
and bamboos being all grown in some numbers.
126
WILLIS AND GARDINER : BOTANY
A few plants were collected on Landu and Mafaro islands,
but only such as had apparently not been seen, or rarely
been seen, elsewhere.
Tiladummati Atoll.
From the information I collected this seems to resemble
Miladumadulu, but the atoll is famous for its plantains,
which are exported even as far as Male. Pandanus is much
grown for sails, and Hibiscus tiliaceus for ropes. The
people of this atoll formerly paid annual visits to the Malabar
and South-West Indian coast. They are a very enterprising
set, and have probably introduced a large number of the
cultivated plants of the northern part of the group. Even
now they regularly bring fresh strains of plantains from
Ceylon and Southern India, as they say that their own
deteriorate. One of my boys, too, from Nolewangfaro,
collected quite a large number of seeds in Addu and Male,
which he intended to try when he got home.
IV.~TRE MALDIVIAN NAMES OF PLANTS.
In the following table are set out in alphabetical order all
the names known or supposed to be applied to plants by the
Maldivians, with the plants to which they refer. Many of
the names are no doubt wrongly applied, especially those
given by other writers than Ibrahim Didi and J. S. Gardiner.
It is remarkable how nearly all the plants of the flora have
received names ; this is partly to be accounted for, no doubt,
by the smallness of the flora, but suggests invention on the
part of the inhabitants when questioned about the plants,
such as is so troublesome to the botanical collector in Ceylon
or India. The spelling of the names is given as originally
written by the collector. Under the individual species
above only recognized names and spellings adopted by
OF THE MALDIVE ISLANDS.
127
Mr. Gardiner or Ibrahim Didi are mentioned. In this list
other names and spellings are also included.
Ahi, Morinda citrifolia
Ala, Colocasia Antiquorum
Amanaka, Ricinus communis
Anana, Ananas sativus
Annaru, Punica Granatum
Anona, Anona muricata
Asaruma, Mirabilis Jalapa
Asuraffole rai (ma), ditto
re, ditto
hudu (ma), ditto
Ata, Anona squamosa, muricata
Babukeyo, Ai’tocarpus incisa
Bai, Eleusine Corocana
Bambakeyo, Artocarpus incisa
Bambara, unknown
Bambukea, Artocarpus incisa
Banke, ditto
Bari, Solanum Melongena
Barubo, Cucurbita Pepo
Beem magu, Oldenlandia
biflora
Beembi, Eleusine Coracana
Berebedi, Erythrina indica
Bes gobili. Acacia Farnesiana,
Parkia ?
Bilehhui, Ischæmum ciliare
Bilei, Piper Betle
Bilekkattala, Dioscorea, spp.
Bilimagu, Averrhoa Bilimbi
Bilinhui (Bell), unknown grass
Billaton, Piper Betle (leaf)
Billi, Piper Betle
Billikatella, Dioscorea alata
Bimbi, Eleusine Coracana
Bissagu, Acalypha fallax
Bo, Cocos nucifera
Bodu faru, Bryophyllum caly-
cinum
Bodukafa, Gossypium barba-
dense
Bodu limbo, Citrus Medica
Bodu mirihi, Tithonia diversifolia
Boi, Ficus religiosa"^
Bokeo, Pandanus, sp.
Borhi, Tournefortia argentea
Bulukiya, Corchorus capsularis
Burandagondi, Boerhaavia diffusa
Burubo, Cucurbita
Cadapie, Launæa pinnatifida
Cave, Acalypha fallax
Chunpapool, Artabotrys odora-
tissimus
Dagundi kandi, Acalypha pani-
culata
Dai kurandu, Barleria Prionitis
Dakada, unknown.
Dandu filia, Launæa pinnatifida
Dhadukuradi, Pemphis acidula
Dhumburi, Ochrosia borbonica
Diga, Hibiscus tiliaceus
Digga", ditto
Digutiyana, Cassia Sophera
Dikkunfati, Commelina, sp. ?
Divihe keeva, Pandanus, sp.
Diyamudoli, Commelina bengha-
lensis
Dorn (or Don) Moosa (or
Moussa), Allophylus Cobbe
Dorn velam buli, Cassytha fili-
formis
Dommadu, Terminalis Catappa
Donala, Zea Mais
Douera, ditto
Dongkeo, Musa Sapientum
Duburi (Bell), Calophyllum
Inophyllum
Dugajde, Clerodendj-on inerme
Dugeti, ditto
Dumbu, Ficus infectoria
Dumburi, Ochrosia borbonica
Dumpai, Nicotiana Tabacum
Dunnika, Yitex Negundo
* Always called Boi-gas (gas = tree).
128
WILLIS AND GARDINER : BOTANY
Emboo, Glochidion littorale
Emmuli, Oldenlandia umbel-
lata
Eyaganawatura, Oldenlandia bi-
flora
Faga, Momordica Charantia
Falo, Carica Papaya
Fatangu, Cæsalpinia pulcher-
rima
Fatunfaifila, Bryophyllum caly-
cinum
Femfora, Ai-eca catechu
Feru, Psidium Guyava
Fesko, Tephrosia tenuis
Fili, Ficus, sp. ?
Fini fen ma, Rosa, sp.
Fo, Areca Catechu (nut)
Foh, Adansonia digitata
Foni loli, Adenostemma visco-
sum
Foni tula, Ocimum gratissimum
Fuakon, Araca Catechu (nut)
Fufu, Benincasa cerifera ?
Funa, Calophyllum Inophyllum
Gadacolie, Ocimum basilicum
Gai kehabuli, Ocimum sanctum
Gakehebiili, Fleurya interrupta
Gandakoli, Ocimum basilicum
Garada, Abroma augusta
Gelavalie, Premna integrifolia
Gethawcoley, Corchorus acutan-
gulus
Gitakoli, Pouzolzia indica
Giteyokoli, ditto
Gobili, Phyllanthus distichus
Gobu, Terminalia Catappa
Gokkuladuru, Cyperus, sp.
Gulhajaru, Pancratium, sp. ?
Gulu Sampa, Plumeria acuti-
folia
Hai kurudo, Barleria Prionitis
Hala veli, Suriana maritima
Halia veli, ditto
Hande riemo, Cyanotis, sp. ?
Handu, Oryza sativa
Hau (Bell) ^ Kuna ?
Heena, Lawsonia inermis
Heith thala, Tacca pinnatifida
Hekoopie, Desmodium triflorum
Helebeli, Tamarindus indica
Hikundi, Murraya exotica
Himeri, Phaseolus lunatus, and
next
Himerri, Dolichos Lablab, and
last :
Hinbatu, Thevetia neriifolia ?
Hiridigga, Corchorus acutangulus
Hirikulla, Emilia sonchifolia
Hirundu, Thespesia populnea
Hitala, Tacca pinnatifida
Hittala, ditto (Trimen, Dios-
corea)
Hiti, Azadirachta indica
Hondikunavoo, Nephrolepis
tuberosa
Huanduma, Jasminum grandi-
florum
Hudufipila, Aerua lanata
Hudufupila, ditto
Hudu ruva, Calotropis gigantea
Hui, grasses generally
Hui ru, Mariscus Dregeanus
Hunigundifila, Lippia nodiflora
Hunigunditila, ditto
Huvaduka, Plectranthus ?
Huvadukotun, Plectranthus ?
Huvaduma, Jasminum grandi-
florum
Huvago, Acorus Calamus
Huwanduma, Jasminum gran-
diflorum ,
Innapa (Pyrard), Lawsonia
inermis
Irudema, Jasminum Sambac
Iroudemaus (Pyrard), ditto
Jamburool, Eugenia, sp.
Jumbu, Eugenia, sp.
Junapa (Trimen), misprint for
Innapa
OF THE MALDIVE ISLANDS.
129
Kadapie, Launæa piiinatifida
Kadu, Hernandia peltata
Kadu, Cucurbita, sp. ?
Kaduru, Corchorus acutangulus
Kafa, G-ossypium, spp.
Kairi, Cocos nucifera
Kala kiru, Pedilanthus tithy-
maloides
Kalandurugundi, M ariscus Dre-
geanus
Kalukadili, Eclipta alba
Kamaraga, Averrhoa Caram-
bola
Kambulichi, Yernonia cinerea
Kanditoli, Poinciana regia
Kando, Hernandia peltata
Kandu, a Mangrove, see Rhizo-
phora, Bruguiera, Lumnit-
zera
Kandufa, a mangrove swamp.
Kani, Cordia subcordata
Kara, Solanum Melongena,
Citrullus Colocynthis ?
Kandolu, Rhoeo discolor
Karhi, Cocos nucifera
Karhi leibu, Achyranthes
aspera
Karhikeo, Pandanus, sp.
Karikeo, ditto
Karikuburu, Cæsalpinia Bondu-
cella
Kasinji, Andropogon Nardus
Katella, Dioscorea, spp.
Kattala, Dioscorea and Ipomœa,
spp.
Kave, Acalypha fallax
Keembi, Barringtonia speciosa
Kees fila, Kephrolepis exaltata
Keeva, Pandanus, sp.
Kekuri, Cucumis sativus
Keo, Musa sapientum
Kerutina, Euphorbia hyperici-
folia
Khari, Cocos nucifera
Kimbi, Barringtonia speciosa
Kirutina, Euphorbia, spp.
Kochchefai, Agératum cony-
zoides
Konara, Zizyphus Jujuba
Kuda falo, Carica Papaya
Kuda limbo (lubo), Triphasia
trifoliata
Kude, Premna integrifolia
Kudehi, Ficus, sp. ?
Kudibai, Panicum miliaceum
Kudima, Jasminum auriculatum
Kudingaybelamaw, Oldenlandia
biflora
Kudingke, Phyllanthus maderas-
patensis
Kudiraimaveyo, Ipomœa Qua-
moclit
Kudiruvali, Dodonæa viscosa
Kiihada, Cassia occidentalis
Kukulufaifila, Hibiscus Solandra
Kulitula, Ocimum sanctum
Kullafila, Launæa pinnatifida
Kullava, Sonneratia acida
Kulowa, do.
Kuna, Pycreus polystachyus
Kunnaru, Zizyphus Jujuba
Kuredi, Pemphis acidula
Kuradu do.
Kurifila, Ipomœa Turpethum
Kurra, Citrullus, sp.
Kurukuru, Dioscorea fasciculata
Kurumba, Cocos nucifera
Ladu, Mimosa pudica
Lagundi, Commelina, Gloriosa
Lami, Eugenia Jambolana
Lancimo, Andropogon squar-
rosus
Limbo, Citrus Medica
Los, Pisonia morindæfolia
Lunsimoo, Andropogon squar-
rosus
Ma ala, Alocasia indica
Mabori, Tournefortia argentea
Mabula, Abutilon indicum
Mabulan, Sida humilis
Abutilon indicum
130
WILLIS AND GARDINER ; BOTANY
Magu, Scævola Koenigii
Makahikeyo, Pandanus, sp. ?
Makumburuvani, unknown
Makunu hungal, Nephrolepis
exaltata
Makunu fila, Portulaca quadrifida
Mala embu, Stachytarpheta indica
Ma lebu, Trichodesma zeylani-
cum
Maliku ruva, Vinca rosea
Manila, Canavalia lineata
Mapijja, Sesuvium Portulacas-
trum
Maskando, Hernandia peltata
Maskota, Anisomeles ovata
Massagu, Amarantus viridis
Mathewa, Dioscorea, sp.?
Mativa, Dioscorea, sp. ?
Ma tumba, Barleria Prionitis
Medili, Terminalia Catappa
Medukeo, Pandanus, sp.
Medu mokaneo, Pandanus, sp. ?
Meia limbo, Evolvulus, Phyllan-
thus
Meia sagu, Acalypha paniculata
Midili, Terminalia Catappa
Mirajje sai. Sida carpinifolia
kochchefai, Artemisia
vulgaris
Mirihi, Wedelia biflora, and see
bodu-, veo-, walu-, mirihi
Mirus, Capsicum minimum
Mita bodi, Ruellia ringens
Moli, Citrus Aurantium
Munima, Mimusops Elengi
Muraga, Moringa pterygosperma
Muraki, Physalis minima
Muranga, Moringa pterygos-
* perma
Muskotan, Anisomeles ovata
Muskundu, Hernandia peltata
Nagukandi, Hibiscus Solandra
Nanubeddi, Corchorus, Boerha-
avia
Nargis, Pancratium, sp.?
Nika, Ficus, sp. ?
Nita bodi, Ruellia ringens
Nitu badi, ditto
Niyaduru, Citrus Decumana
Nunnay, Stachytarpheta indica
Okkafa, Gossypium herbaceum
Oludukattala, Ipomœa Batatas
Ona hui, Oplismenus, Apluda,
sp.
Orhani, Datura fastuosa
Prumo, Psilotum triquetrum
Rabeburi, Cleome viscosa
Raidadi sagu, Amarantus cau-
datus
Ragundi, Cordia, sp. ?
Raigeda, Portulaca oleracea
Ranarua, Cassia, sp.
Ranaura, ditto
Ranawia, ditto
Ra rohi, Colubrina asiatica
Ra ruhi, ditto
Re asaruma. Mirabilis Jalapa
Re irudema, J asminum Sambac
Re kandolu, Rhœo discolor
Revi, Brassica juncea
Ribu fila, Portulaca tuberosa
Riindu filia, ditto
Rua, Calotropis gigantea
Ruva, ditto
Sabudeli, Chrysophyllum, sp.?
Sagu, Amarantus, Acalypha
Sakkeyo, Artocarpus integrifolia
Sannipoo, Eragrostis plumosa
Satavaru, Asparagus racemosus
Sembu nerinchi, Tribulus ter-
restris
Semper Beddha, Plumeria acuti-
folia
Taburu, Ipomœa bilöba
Talafuri, Canavalia ensiformis
Tavakarhi, Lodoicea Sechellarum
Thaburu, Ipomœa biloba
OF THE MALDTVE ISLANDS.
131
Tilo lagundi, Commelina Kurzii
Timbi, Barringtonia speciosa
Tinfaikattala, Dioscorea penta.
phylla
Tora, Luffa ægyptiaca ?
Tubuli hui, Eragrostis plumosa
Udandi, Saccharum officinarum
Uni, Guettarda speciosa
Bambusa, sp. ?
Ura, Setaria italica
Vaffufuli, Acalypha indica
Yakkunfati, Commelina, sp. ?
Yarukundu, Trewia nudiflora
Yele buli, Cassytha filiformis
Yeo falo, Carica Papaya
magu, E volvulus alsinoides
mirihi. Convolvulus, Leucas
Yeppila, Herpestis Monniera
Yeyodigga, Sida humilis
Yihafilia, Indigofera tinctoria ?
Yihagiguni, Crotalaria retusa
Yihalagondi, Gloriosa superba
Yiyula tari, ditto
Yihatoli, Pachyrhizus angu-
latus
Walu kafa, Yernonia cinerea
mirihi. Convolvulus parvi-
florus
Wang, Piper Betle
Wihelia gundi, Gloriosa superba
Zaggumu, Argemone mexicana
Zoowari, Andropogon Sorghum
The meanings of some of the commoner words above are
given in the following list : —
Ali, ash gray
Bodu, big
Buli, fish-hook
Dambu, dark
Digu, long
Dorn, light or white
Fai, ankle, leg
Faro, a ring-shaped reef awash
Feli, cloth
Filia, thread (cotton)
Fehi, green
Fune, deep
Gas, tree
Gundi, stool
Hini, cold
Hudu, white
Huni, hot
Iru, sun
Kafa, a cloth colour
Kalu, black
Kandi, deep sea, deep passage
Kuda, small
Kuding, child
Ma (suffix), very ; flower ?
Magu, path
Noo, blue
Be, red
Riindu, yellow
Tilia, dirty
Tiri, low
I Yeli, sand
I Walu, a well
V,—THE ECONOMIC PRODUCTS OF THE
MALDIVES.
Under this head must be considered not only those plants
which have certainly been introduced for purposes of
cultivation, but also those which are known, though wild, to
(18)
132
WILLIS AND GARDINER : BOTANY
yield useful products. All the plants of the former class are
printed in capitals in the list of the flora. Some plants, such
as the cocoanut, though now cultivated, have probably been
introduced by natural agencies. Many of the cultivated
plants in the list have without doubt been introduced quite
recently by Ibrahim Didi and others, and are as yet hardly
to be found in the outlying islands. There has been a
marked increase in the successful acclimatization of food
and other plants in the last twenty years. This is largely
due to Ibrahim Didi, Dorimenokiligefanu, late Chief Vizier,
to whom the people owe a very great debt in this respect.
The Sultans, too, have taken more interest in the matter, and
the nobles at Male at present vie with one another in
the rare (^.e., with regard to the Maldives) plants and shrubs
growing in their compounds.
The effect of the eruption of Krakatoa in 1883 has not
been sufficiently emphasized. Before then the islands had
no pumice, and it was not possible to grow a large number of
the food plants in many of the islands. Now plantains,
&c., grow almost anywhere, a basket of pumice being placed
around the roots of each.
The certainly introduced and cultivated plants of the
Maldives in the list above given number (including doubtful
identifications) 98, belonging to 73 genera and 33 families.
Sixty-one genera and 12 families are represented by culti-
vated plants only.
We may classify the economic products into groups as
follows, using the classification of the “ Indian Agricultural
Ledger ” -
Group I.— Gums^ Resins, &c.
There are several trees in the flora which are elsewhere
used for the source of gums and resins, but we have no
record of any being used by the Maldivians. Such are
Azadirachta, Acacia Farnesiana, Bread-fruit, Mango, Moringa.
The saps of bread-fruit, mangrove, and Barri ngtonia are
OF THE MALDIVB ISLANDS.
1?>Z
used for caulking boats with coir or native cotton, the whole
mixed with soot of cocoanut husks or preferably of Barring-
tonia fruits.
Group II. — Oils^ Perfumes, &c.
The chief oil used in the islands is of course that of
the cocoanut. Castor oil is sometimes used as a purgative.
Calophyllum Inophyllum oil is used as an ointment. The
Margosa has lately been introduced.
Perfumes are obtained from the flowers of the Jasmine,
Plumeria, rose, basil, and probably others, such as Guettarda,
&c., and also from the Khus-khus and Citronella grasses.
Group III. — Dyes and Tanning Stuffs.
The principal source of dye seems to be the root of
the Ahi, Morinda citrifolia, from which they prepare the
red dye used for mats, &c., by mixing a decoction with lime.
A black dye, sometimes used in the mats, is made by boiling
gallnuts and rusty iron together in cocoanut water (Bell),
but the deepest black is made from plantain sap. A deep
black is made from Mangrove mud, and lighter shades
to brown from Mangrove sap. Black is also obtained by
mixing cocoanut shell charcoal with oil. Lacquer work
is carried on only in Turadu (S. Mahlos) ; the lac and
colours used are obtained from Calcutta. The flora also
includes Indigo and the Chay root (Oldenlaudia umbellata),
but these are not now used as dyes. Henna is not
uncommon. Of tanning substances there appear to be none
in use, fish skins, &c., being simply stretched and dried ;
there are several possible sources, such as Mangroves,
Cassias, &c. [The only mammals are a bat, a rat, and a
mouse, with a few semi-wild white rabbits,]
Group IV. — Fibres.
The chief fibre in use is of course coir. Like the Lacca-
dives, the islands produce a very fine quality of this fibre,
light in colour, fine, and strong, which brings a high price in
134
WILLIS AND GARDINER : BOTANY
the market, and is exported in large quantity, Tte best
quality is made in Addu Atoll, where the fibre made is
longer.* The fibre of Hibiscus tiliaceus, the Diga, and of
Hernandia peltata, also bread-fruit sometimes, is used for
ropes and lines and for sails. Cotton is cultivated to a slight
extent, and a certain amount of cloth is made in Mahlos and
Addu, and perhaps elsewhere. Several other fibre-yielding
plants, such as Corchorus, Calotropis, &c., exist in the fiora,
but we have no information as to their actual use, if any. In
the southern atolls a great industry is carried on in the
making of the fine Maldivian mats, of their kind about the
finest in the world, beautifully woven, in tasteful patterns,
and dyed with very fast colours, black, white, and brown-
yellow. The chief source of material is the kuna, Pycreus
polystachyus, a common sedge in Ceylon, where however it
does not appear to be utilized. Fimbristylis spathacea is
also used. These mats were formerly used for sarongs.
The weaving of cloth in the Maldives may have been derived
from that of these mats.
The leaves of Pandanus {q.v,) are used for mat- and
sail-making.
Group V.— Drugs.
The only known plants of the fiora from which drugs are
recorded as used in the above list are Calotropis, Ricinus,
Calophyllum, and Rhoeo, but probably many others, e.g,^
Croton, Azadirachta, Vitex, are used. “ Internal medicines
are very little used, being regarded as contrary to the Koran.
(Allah willed the sickness, perhaps as a punishment for sin,
and it is useless to strive against it.) There has always been
a small class of witch doctors (men), who prepare draughts,
but these are mere decoctions over which a charm has been
wrought, and not medicines. More definite ideas prevail in
the southern atolls, particularly Suvadiva, where I have no
doubt that all the above and many more are used.” — J. S. G.
* I understood that an especially long nut was grown in Addu for
this purpose ; I saw some which appeared to be merely selected examples
of the Khari, but my information would lead me to believe that there is an
especial variety which gives little oil but excellent coir. — J. S. G.
OF THE MALDIVB ISLANDS.
135
Group VI.— Foods and Edible Products.
Taking first the cereals and other staples, we find that the
bulk- of the rice used is imported from Bengal. Bimbi, Eleu-
sine Coracana, is a good deal cultivated, and occasionally some
of the millets, Panicum miliaceum and Setaria italica, and a
little maize. The cocoanut, as in other tropical countries,
provides many of the necessaries of life. Of root crops, half
a dozen kinds of true yams (Dioscorea) are grown, as well
as the sweet potato, the cassava or tapioca, and Colocasia
and Alocasia. The root of Tacca furnishes food in times of
scarcity. Of pulses, there are Dolichos Lablab, Pachyrhizus
angulatus, Phaseolus lunatus, and Canavalia ensiformis, all
common beans of Ceylon and Indian cultivation. Mangrove
rootlets are often eaten.
Of spices, arecanut is both cultivated and imported, betel,
pepper, and chillies are grown, and a little black pepper.
Murraya and Brassica are recent introductions.
Sugar cane is cultivated to a slight extent, and some of the
Amaranti are perhaps used as potherbs.- Tobacco is occa-
sionally cultivated.
Of fruits, eaten raw or in curries, there is a fairly long list.
The most important are in order : the papaw, bread-fruit,
pumpkin, plantain, pomegranate, lime. Orange, lemon, shad-
dock, watermelon, soursop, guava, cucumber, piueapple, &c.,
are also frequent, and custard apple, carambola, blimbing,
jambu, jak, brinjal, mango, &c., have been introduced. The
wild or semi-cultivated fruits of Triphasia, Zizyphus, Allo-
phylus ? Tamarind, Terminalia, Sonneratia, Physalis, Panda-
nus are also eaten. The cocoanut has already been men-
tioned.
Murraya (rare) and Moringa are cultivated as curry stuffs,
and Premna is perhaps also used. Coffee and sugar are
imported, but sugar is principally made from cocoanut sap
(ra). No alcoholic drinks are known. Coffee is used by
everyone who can afford it, tea by only a select few who have
learnt the habit in India or Ceylon.
136
WILLIS AND GARDINER : BOTANY
Group VII. — Timbers.
Timber is not plentiful in the islands, and some is at times
imported. Practically every possible tree is used for timber.
Kuredi (Pemphis) is used principally for firing. Good houses,
boats, and other things are constructed from the cocoanut
wood. Pandanus is used in a temporary house. Flooring is
made of any light wood. Among the list of plants in the
islands given above there occur the following which have
more or less useful timber : Gyrocarpus, Barringtonia,
Scævola, Chrysophyllum ? Mimusops, Cordia, Premna, Vitex,
Pisonia, Hernandia, Ficus spp., Artocarpus spp.. Bamboo.
Ornamental Plants.
Pyrard de Laval records the islanders as very fond of bright
and sweet-smelling flowers,^ but the number upon the list
above is very small, and there are many which one would
expect to have found which are absent from the list. To
enumerate them in systematic order, there are : (i.) common
everywhere, Cassia Sophera (possibly not cult.), rose, Guet-
tarda speciosa (wild). Jasmines, Vinca rosea. Plumeria acuti-
folia, Ocimum spp.. Mirabilis Jalapa, &c. ; (ii.) occasionally
cultivated, Artabotrys ? Gynandropsis (wild). Hibiscus Rosa-
sinensis, H. Abelmoschus ? Abroma augusta, Clitoria, Cæ-
salpinia pulcherrima, Poinciana, Acacia Farnesiana, Passi-
flora cœrulea, Panax, sp. ? Aralia Guilfoylei, Tithonia diversi-
folia, Thevetia, Ipomœa Quamoclit, Datura suaveolens,
Clerodendron fragrans, Antigonon Leptopus, Bougainvillæa
spectabilis, Celosia, Amarantus caudatus, and others. Many
are quite recent introductions and found only in Malé Atoll.
For a people fond of flowers, the list, even considering the
unfavourableness of the soil and climate to gardening, is a
small one.
* They certainly care for the scent, not the prettiness, now. I doubt
Pyrard here, and think he must have g-ot confused with his later experi-
ences in India. There is nothing to show that the people ever attained
this pitch of civilization. — J. S. G.
OF THE MALDIVE ISLANDS.
137
Comparing the list of cultivated plants in the Maldives with
that of the Laccadives and Minikoi,"^ we find that it is much
larger, as might be expected from the greater size, wealth,
&c., of the group, containing at least 46 species that do not
occur in the northern islands. It contains about 98 species,
as against about 58. Among the 12 species found in the
northern group not cultivated in the Maldives are Annatto,
Ægle Marmelos, Ground-nut, Cowpea, Physalis peruviana,
Phyllanthus Einblica, Agave vivipara, and a yam. Among
the Maldivian plants not cultivated in the Laccadives or
Minikoi are Custard apple, Carambola, Blimbing, Zizyphus,
Jujuba Roses ? Plumeria, Brinjal, Tobacco, Cassava, Ficus
religiosa. Pineapple, some yams, &c. Some of these, e.g,^
Zizyphus, probably were introduced from the West, whence
much of the native stock came to the islands, others, e.g.^
Plumeria and Ficus religiosa, probably owe their introduc-
tion to pre-Islamic times, prior to the arrival of Arab, Malay,
and possibly Persian settlers.!
VL—THE ORIGIN OF THE FLORA.
The flora of the islands is undoubtedly recent and derived
from abroad. We must now proceed to deal with the
question of the sources from which it was derived, and the
manner in which it reached the islands.
Plants introdyced unintentionally by Human
Agency.
The paths of human migration and trade, all over the
world, are marked by weeds that have accompanied the
wanderers or been carried in the articles of trade, uninten-
tionally, but none the less certainly. Going over the above
list of the flora there can be little doubt that at any rate the
* This may be partly due to the hurricanes which at times sweep
the latter islands, destroying and washing awaj' many plants.
t There is no evidence (sensu stricto) of Buddhism, though there is of
Hinduism, in the Maldives. Cf. Gardiner : The Natives of the Maldives,
Proc. Camb. Phil. Soc., XI., 1900, p. 17.
138
WILLIS AND GARDINER ; BOTANY
species mentioned in the following list have reached the
Maldives in this manner. Many probably have come in the
bags of imported rice, others attached to articles of clothing
or to packages, others in the earth attached to the roots of
imported plantains and other useful plants, and in similar
ways dependent on man. Species in brackets occur in the
Laccadives, Minikoi, or Diego Garcia, but not in the
Maldives : —
Argemone mexicana
Cleome viscosa
Gynandropsis pentaphylla
Polygala erioptera
Sida humilis
Abutilon indicum
Hibiscus Solandra
Corchorus acutangulus
Tribulus terrestris
Crotalaria retusa
( verrucosa)
(Indigofera cordifolia)
Cassia occidentalis
Tora
glauca
auriculata
Mimosa pudica
(Passiflora suberosa)
Spermacoce ocymoides
Hewittia bicolor
Evolvulus alsinoides
(Linaria ramosissima)
Ruellia ringens
Justicia procumbens
Rungia parviflora
( linifolia)
(Peristrophe bicalyculata)
Anisomeles ovata
Leucas biflora
zeylanica
( aspera)
Amarantus spinosus
Euphorbia hypericifolia
Phyllanthus maderaspatensis
Urinaria
Niruri
( rotundifolius)
(Claoxylon Mercurialis)
Acalypha paniculata
indica
fallax
Fleurya interrupta
Pouzolzia indica
(Aneilema ovalifolium)
Commelina benghalensis
Kurzii
Cyanotis cristata
(Oplismenus Burmanni)
compositus
(Setaria verticillata)
Ischæmum ciliare
Apluda varia
(Cynodon Dactylon)
(Andropogon contortus)
Eleusine ægyptiaca
( indica)
A total of 56 species, of which 41 occur in the Maldives
proper. They represent 41 genera and 18 families, of which
28 genera and 4 families are represented in this list alone,
taking all the islands, or 23 genera and 4 families of the
Maidive flora only.
OF THE MALDIVE ISLANDS.
139
Now let us examine into the local distribution of these
species, as given in detail above. Dividing the islands into
the following groups : The Chagos, Southern Maldives (Addu
and Suvadiva Atolls), Central Maldives (Haldumati to Male
Atolls, south of the Kardiva Channel), Northern Maldives
(north of the Kardiva Channel) ,Minikoi, Western Laccadives
(the chain including Akati and Bitrapar), Central Laccadives
(Kiltan, Kadamum, Ameni, &c.), and Eastern Laccadives
(Anderut, Kalpeni, &c.) — we have —
Chagos.
S.M. C. M.
N.M.
Mkoi. W. L.
G.L.
Species cert. 2 ...
12 .,
.. 38 .
.. 26 .
.. 21 ... 21 .
.. 30
prob. — ...
14 .,
.. 2 .
.. 7 ..
.. — ... — ...
.. —
Total ... 2
26
40
33
21 21
30
In the first line are given only those species of which we
have a certified record of occurrence ; in the second are given,
for the Maldives only, those species which are so common
elsewhere in the group that their occurrence is probable,
e.g.^ Abutilon indicum. Cassia Tora, Evolvulus alsinoides,
Fleurya interrupta, Ischæmum ciliare, for the Southern
Maldives. The figures for the Chagos are certainly incom-
plete, owing to Mr. Hemsley having excluded this class of
plants as far as possible. If we examine the remainder, we
arrive at the following conclusions : — The greatest number of
weeds of this class is to be found at the centres of maximum
traffic and commerce, i.e., the central islands of each group.
Many of the 40 species recorded for the Central Maldives are
found in Male or Malé and Hulule (close by) only, and the
number decreases as we pass to the outlying islands. In the
Eastern Laccadives, though so near to the Indian coast, we
find only a few species. Male is not the only point to which
traffic comes : the northern and southern atolls (see above,
Tiladummati, &c.) have a certain amount of direct communi-
cation with India, which has caused introduction of some,
weeds not known in Male. Leaving Diego Garcia out of
consideration, in which the weeds are of Seychelles or
(19)
140
WILLIS AND GARDINER ; BOTANY
cosmopolitan types, we find that the weeds are, as a rule, the
commonest weeds of India and Ceylon. Poygala erioptera
is the only Maldivian weed not known in Ceylon, hut is
common in India. There is a considerable direct trade
between Malé and Calcutta, to which is probably due the
introduction of many Indian weeds.
Taking Maldives and Laccadives together, the commonest
and most widely distributed of these weeds are probably Sida
humilis, Abutilon indicum, Anisomeles ovata, Phyllanthus
maderaspatensis, P. Niruri, Acalypha indica, A. fallax, and
Oplismenus compositus, each of which occurs in at least six
of the seven divisions made above.
The number of these weeds is greater on the more
cultivated islands ; their presence is largely dependent on
cultivation. There are only a few of them on the wilder
islands where there is less traffic, less cultivated ground, and
greater competition with other plants. Most of the species
occur in more than one of the divisions, and are often
distributed in such a way as to indicate having spread from
one spot in each archipelago or group. All the Maldivian
weeds occur in Malé or Goifurfehendu Atolls, with the
exception of Hewittia bicolor, recorded only from Mahlos.
The following islands have fairly complete floras, and show
the following numbers of introduced weeds of this list :
Male 27 at least. Hulule 13, Hedufuri (S. Mahlos) 16,
Goifurfehendu Atoll 17, Veimandu (Kolumadulu) 13.
Species probably introduced by Birds.
All previous investigations tend to show that the agency
of birds in providing distant islands with plants is a com-
paratively unimportant one, so far as number of species is
concerned. In the case of plants with fleshy fruits, whose
seeds are afterwards dropped, it is evident that in general
only islands lying at comparatively short distances can be
reached, but in the case of fruits or seeds sticking by means
of hooks or gum, or enclosed in mud pellets attached to the
OF THE MALDIVE ISLANDS.
141
feet, greater distances may be traversed.* Let us take the
fleshy-fruited plants first. Of these we have the following : —
Ardisia humilis
Of these 25, it is all but certain, as we have seen above, that
Zizyphus, Allophylus, Physalis have been intentionally
introduced for cultivation. The occurrence of Vitis Linnæi
is very doubtful. Colubrina, Sonneratia, Morinda, Cassytha,
Premna, and Ardisia are littorals which may be introduced
by sea currents, though when once introduced they may be
further spread by birds. Datura fastuosa and Solanum
torvum are common cultivation weeds. The Asparagus is
almost certainly intentionally introduced, as it only occurs
in Male, and has the same name as in India and Ceylon, and
there are doubts about Vitex and the Fici. We can thus
only accept as almost certainly introduced by means of
seed-dropping by birds the following species : — Flacourtia,
Pleurostylia, Vitis two spp., Cephalandra, Pavetta, Lantana,
seven in all, belonging to six genera and six families. Of
these, five occur only in the Eastern and Central Laccadives
near to the mainland of India, one occurs only in the
Western Laccadives, one only in the Northern Maldives.
We may therefore reasonably conclude (i.) that there is a
fair amount of evidence that plants may be carried by birds
in their intestinal canals from India to the nearer Laccadives,
a distance of 100-150 miles, and possibly from Ceylon to
the Maldives, a distance of 350-400 miles ; if this occurred,
it must have been in the north-east monsoon, when the
(Flacourtia Sepiaria)
(Pleurostylia Wightii)
Zizyphus Jujuba
Colubrina asiatica
Vitis Linnæi
( quadrangularis)
( carnosa)
Allophylus Cobbe)
Sonneratia acida
(Cephalandra indica)
(Pavetta indica)
Morinda citrifolia
(Solanum torvum)
Physalis minima
Datura fastuosa
Lantana mixta
Premna integrifoha
Vitex Negundo
Cassytha filiformis
Ficus benghalensis
Asparagus racemosus
infectoria
Retusa
sp.
142
WILLIS AND GAB DINER : BOTANY
wind may travel in the necessary direction 250 or more
miles in the day ; (ii.) that there is no evidence to show that
this method of transport has certainly carried plants over
wider distances than 400 miles and probably even 350 ;
(iii.) that when once introduced by any means, plants with
fleshy fruits may very probably be largely distributed
among the islands by birds, many of the fleshy-fruited
species above mentioned being very widely scattered among
the islands.
Pass on now to the carriage of seeds attached by claws
or hooks to the feathers of birds. The evidence -that this
carriage ever takes place in nature is small, but there are
a few known probable cases. In the flora above enumerated,
the following are possibilities : —
Sida carpinifolia
(Urena sinuata)
Desmodium triflorum
Adenostemma viscosum
(Bidens pilosa)
(Plumbago zeylanica)
Boerhaavia diffusa
Pisonia aculeata
morindæfolia
Achyranthes aspera
Any or all of these may obviously just as likely have been
introduced by man, attached to clothing or other articles.
Taking the list for what it is worth, we find that Sida and the
Pisoniæ occur in the Maldives only, Bidens, Urena, and
Plumbago in the nearer Laccadives, and the others through
both groups. Boerhaavia and Pisonia are very probably sea-
borne. So far as this evidence goes, therefore, it shows the
same results as that of the dropped seeds.
Now, take the case of the plants with seeds or fruits so
small that they may be carried in mud pellets attached to
the feet of birds. This evidently applies mainly to the plants
of wet places with muddy ground, not so much to sandy soil.
We thus get the following list : —
Portulaea oleraceaa
quadrifida
tuberosa
(Ammannia baccifera)
-Sesuvium-Portulacastrum
Eelipta alba
Wedelia calendulacea
Herpestis Monniera
(Striga lutea)
Lippia nodiflora
(Polygonum barbatum)
■Euphorbia piluliferaj
OF THE MALDIVE ISLANDS.
148
aS'well as the following possible grasses and sedges
Pycreus polystachyus
(Fimbristylis diphylla)
( glomerata)
Cladium Mariscus
(Oplismenus Burmanni)
compositus
Apluda varia
Eleusine ægyptiaca
Of these 20, other observers have shown that there is a
probability of sea carriage in the case of the Portulacas,
Lippia, and possibly some of the others. Leaving out those
mentioned, we find Ammannia baccifera, Polygonum bar-
batum, and Fimbristylis diphylla confined to the Eastern
Laccadives, F. glomerata to Diego Garcia, while the rest are
mostly widely spread among the archipelagos, except Cla-
dium, which is confined to the extreme south. Cladium
occurs in the African islands to the west, and is not found
in India proper or in Ceylon. We cannot therefore say that
we have any evidence for carriage of seeds by attachment
to feet of birds which is unimpeachable, but we may say
that there is a slight amount of evidence in favour of the
possibility, and that if this evidence be ever proved true, it
supports, as does that of the dropped seeds and the hooked
fruits or seeds, the view that the transport is of few species,
and over comparatively short distances, usually not exceed-
ing 200 miles. The poverty of the islands in land shells is
another argument against extensive transport by birds. The
birds of the group are not such as to lead one to believe that
they ever had any marked influence in bringing plants to
the Maldives. The hooded crow is found in every inhabited
island, and every shore has at times sand-snipe, curlew, and
plover. A rail is found in the densest bush, and seabirds—
terns, boatswains, &c. — are found everywhere. Of migrants,
a few of the swallow family winter in the Maldives, together
with several species of kite and owl, and a roller in consider-
able numbers ; green parrots, too, the only truly gramini-
vorous forms, are occasionally seen.
The flying-foXi Pteropus medius, hangsin enormous num-
bers during the day from nearly every banyan in the group,
144
WILLIS AND GARDINER : BOTANY
and must be of importance in distributing these, and pro-
bably other species, among the islands.
Transport of Seeds, &c., by Wind Agency.
On land and for short distances,* this is one of the most
important of transport agencies, but very little evidence has
as yet been collected to show that it is the means of carrying
many seeds or plants across wide stretches of sea. Looking
through the lists of plants above, the only fairly certain
cases of wind transport are those of the spores of the ferns
and Psiiotum, including the following species : —
(Asplénium æquabile)
( longissimum)
( falcatum)
Nephrolepis exaltata
( tuberosa)
( cordifolia)
Thamnopteris Nidus
Nephrodium molle
( unitum)
(Pteris marginata)
Psiiotum triquetrum
Of these, five, one of which is an endemic species, are
confined to Diego Garcia, one to Minikoi, one to the Eastern
Laccadives, and the others are fairly widely scattered among
the islands. We may thus conclude that there is no distance
by which any of these islands are separated from one another
or from the mainland or other islands, which is too great for
the transport of cryptogamie spores by wind. It must be
remembered that the soil and climate of most of the islands
are not highly favourable to cryptogams. Treub, in his
interesting paper on the new fiora of Krakatoa, has shown
how the ferns are about the first vegetation to appear, and
has drawn from his observations the conclusion that in the
colonization of distant oceanic islands, other than mere
banks or small reefs, with plants, the ferns must take a great
part, and form a large part of the vegetation, conclusions
borne out by the actual composition of the floras of Juan
Fernandez and Ascension. Passing on to the more proble-
matic cases of wind transport, we have the case of the two
* Cf. Willis and Burkill : The Flora of the Pollard Willows near Cam-
bridge, Proc. Camb. Phil. Soc., 1893.
OF THE MALDIVE ISLANDS.
145
Asclepiads, Tylophora asthmatica and Leptadenia reticulata,
found in the Laccadives, and a third, Asclepias curassavica,
in Diego Garcia. The presence of the first, which is widely
spread in the islands, may be due to intentional introduction,
as its root is a valued medicine, used as ipecacuanha, but
there is a chance that the second may have been introduced
by wind, and thus have been carried about 150 miles. Ascle-
pias curassavica may have been carried to Diego Garcia by
wind, but is a cosmopolitan tropical weed wherever cultiva-
tion occurs.
Lastly, there are the pappose Compositæ, but with regard
to all these, which, as a glance at the list will show, are the
commonest of Indian weeds, there is at least as much like-
lihood of accidental introduction through the agency of man,
and no deductions as to wind transport can be based on their
presence, though their commonness in the islands is no
doubt partly due to their easy local transport by wind.
Another argument against any great transport by wind is
the poverty of the insect fauna of the islands. The north-
east monsoon brings a few butterflies and other insects from
India, so that it is possible that some of the parachute seeds
or fruits, e.g., those of Asclepiads or Compositæ, may be
carried over also.
Introduction of Plants by Ocean Currents.
Of all the agencies in the stocking of oceanic islands, this
is the most important, and it has been fully described by
Schimper,* Hemsley,! and others. Many, especially the
most widely distributed, of the littoral plants of the tropics
are provided with admirable adaptations for the floating of
their seeds or fruits over long fstretches of sea. There is no
need to go into details here with regard to these mechanisms.
We shall merely extract from the lists of plants of the
Laccadives, Minikoi, Maldives, and Chagos the species about
* Die Indomalayische Strandflora. Jena, 1891.
f Botany of the Challenger Expedition, Vol. I., 1885.
146
WILLIS AND GARDINER : BOTANY
which there can be practically no doubt that they have been
introduced in this way. These are : —
Scævola Koenigii
A total of 43 species, representing 39 genera and 24 fami-
lies, of which 39 species, 37 genera, and 24 families are
Maldivian.
Of these 43 species, only 36 occur in Ceylon, and only 30
and 29 on the Malabar and Coromandel coasts of Southern
India respectively (or 39 in Ceylon and Southern India
together), so that it is evident that the sea-borne flora of the
islands at any rate is derived in part from other countries
than those nearest to them. We may best consider the
origin of this flora by dealing flrst with the effects likely to
follow from the directions of the currents and winds
at different seasons. Reference to any atlas of physical
geography will show these ; a good current map forms
Plate I. in Schimper’s work, and the map at the end of this
paper gives the currents for the western half of the Indian
Ocean.
Calophyllum Inophyllum
Hibiscus tiliaceus
Thespesia populnea
Suriana maritima
Dodonæa viscosa
Canavalialineata (obtusifolia)
( turgida)
Vigna lutea
Sophora tomentosa
Cæsalpinia Bonducella
(Afzelia bijuga)
Acacia Farnesiana
Rhizophora mucronata
Bruguiera caryophylloides
Terminalia Catappa
Lumnitzera racemosa
Barringtonia speciosa
Pemphis acidula
Guettarda speciosa
Wedelia biflora
Launæa pinnatifida
Ochrosia borbonica
Cordia subcordata
Tournefortia argentea
Ipomœa grandiflora
denticulata
biloba
Clerodendron inerme
Aerua lanata
Hernandia peltata
( ovigera)
Euphorbia Atoto
Agyneia bacciformis
Crinum asiaticum
Pancratium zeylanicum
Gloriosa superba
Pandanus odoratissimus ?
Mariscus Dregeanus
Spinifex squarrosus
Thuarea sarmentosa
Eragrostis tenella
(Lepturus repens)
OF THE MALDIVE ISLANDS.
147
Consider first the wind. South of latitude 10° S. the
south-east trade blows throughout the year. North of this
is the monsoon region, in which from April to September
there blows the south-west monsoon of India and Ceylon,
and from October to March the north-east monsoon. South
of the equator the wind is south-east, at the equator about
south, and north of the equator it is south-west to west.
In the other monsoon the wind is north-east over Ceylon,
northerly nearer to the equator, and north-west to the
south of it.
Now consider the currents. The great equatorial current
fiows all the year round in a westerly direction, passing
through the Malay Archipelago (especially between Java
and New Guinea and Australia), and crossing the Indian
Ocean south of the equator. It splits south of the Chagos
Archipelago, the northern half going round the Seychelles
and returning along the equator to Sumatra. Part of the
southern stream reaches and rejoins this via the Mascarene
Islands and Madagascar. North of this constant stream in
the trade-wind region are the variable currents of the
northern Indian Ocean, which change with the monsoons.
In the south-west monsoon the currents run in general to
the east, in the north-east monsoon to the west, round the
Arabian Sea and the Gulf of Bengal, as shown on .the
maps.
Now consider the probable results to be expected in a
long period of time during which these causes work together.
The equatorial westerly current, aided by the south-east
trade-wind, will carry to the Seychelles all or nearly all
those littoral species of the southern parts of the Malay Archi-
pelago and North Australia which have adaptations good
enough to carry them safely over the long distance. On the
way a few species may be left in the Chagos Archipelago,
but these islands do not lie in the main course of the
current, nor offer any great intercepting area. The return
current will carry the littoral fiora of the Seychelles to some
(20)
148
WILLIS AND GARDINER : BOTANY
extent to the Chagos, and also to Sumatra, whence it may
spread a little into the Malay Archipelago. During the
south-west monsoon the water north of this is also moving
eastwards, and aided by the wind may carry the Seychelles
plants to the Maldives and even to the south of Ceylon,
whence the current flows round the Bay of Bengal to the
Malay Peninsula and the Andamans. Other currents during
this monsoon may carry the plants of the Seychelles by
longer routes to the Laccadives and the Malabar Coast, and
bring with them a few from the coasts of Africa and Arabia.
On the change of the monsoon and the currents, the plants
of the Malay Peninsula may be carried to the Coromandel
Coast, Ceylon, and the Maldives and Laccadives, more
reaching the more southerly islands through which the main
stream goes. The same streams and winds may carry a few
Ceylon and Indian and African forms to the Seychelles.
We may therefore expect, on the whole, to find the
following results after the lapse of a long period of time.
The littoral sea-borne flora of the Seychelles will consist of
a large number of the more southern Malayan forms, with
a sprinkling of species from East Africa, Madagascar, Arabia,
and India. Many of these will also reach Diego Garcia and
the other islands of the Chagos group, whilst owing to the
much greater intercepting area of the Maldives most of the
Seychelles littoral flora will in all probability reach those
islands during the south-west monsoon, which may also
bring a few African or Arabian forms, and during the north-
east monsoon many of the -northern Malayan forms will
reach the east and south of Ceylonand the Maldives, Avhile a
few Ceylon and south-west Indian forms may also be carried
to the latter. The Maldives lying much more in the track
of the currents and nearer to the sources of supply will
have a richer flora than the Laccadives, so far as the littoral
sea-borne plants are concerned, and also, when we consider
the currents and the greater strength of the south-west
monsoon, a richer flora than Ceylon in the Seychellan (i.e,.
OF THE MALDTVE ISLANDS.
149
chiefly’ South Malayan) and even perhaps in the North
Malayan types. If we examine the actual statistics of the
distribution of the flora of the islands, we shall find them in
pretty fair general accordance with these views. It would
lead too far to go into the details of the littoral sea-borne
flora of the Seychelles in this paper, but in general it is of
the Malayan type, especially the southern type, with a
sprinkling of other forms.""
Of the 43 species enumerated above, Afzelia bijuga and
Hernandia ovigera are found only in Diego Garcia ; both of
these are Seychellen plants, and the former at any rate also
Malayan. Of the whole number, 20 are Polynesian, 34 or
more occur in the Malay Archipelago, 36 in the Malay
Peninsula ; of the last-named, 25 also reach the coasts of
Burmah ; 30 occur on the Malabar, 29 on the Coromandel
Coasts of the Indian Peninsula, while 36 occur in Ceylon,
many of them only on the southern coasts ; the Maldives
have 39, Minikoi 29, the Laccadives 22 ; 32 or perhaps 34
occur in the Seychelles, Madagascar, and Mascarene Islands,
14 in Diego Garcia.
All the Diego Garcia species occur in the African Islands,
and all, except perhaps Hernandia ovigera, are Malayan ;
there is no difficulty in accounting for the origin of the flora
on the views given above as to the distribution of plants by
the currents and winds. Turning to the Maldives, and
dividing the group as before into South, Central, and
North Maldives, we have in each of these about 37 species,
and in the whole Archipelago 39. The species peculiar to
the South Maldives are Agyneia baceiformis. Pancratium
zeylanicum, and perhaps Lumnitzera racemosa, the first oc-
curring in the African Islands, Malay Peninsula, and Ceylon,
^ I find in my notes that I saw three Ceylon or Malay outrig-ger boats
cast np, one buoy, and one Malay pran. I also passed a catamaran when
going’ down in October, about 30 miles east of Fadiffolu Atoll. — J. S. G-.
The well known double cocoanut or Coco-de-mer of the Seychelles was
formerly called Cocos maldivica, its^f ruits being' so frequently cast ashore in
the Maldives.
150
WILLIS AND GARDINER : BOTANY
SO that its actual immediate source can hardly be determined;
the second occurring in the Malayan region and Ceylon, but
not in the African Islands, and turning up also in Minikoi ;
while the third occurs in the Malay Peninsula, Ceylon, and
East Africa, though not in the African Islands. In the
Central and Northern Maldives, though apparently not in
the Southern, we find Canavalia lineata (Malaya, African
Islands, Ceylon) and Gloriosa superba (Malaya, Eastern
Africa, Ceylon). The whole group of Maldives contains four
species, or including the Chagos, Laccadives, and Minikoi,
seven species not found in Ceylon, but all of which occur
in the African Islands and the Malayan region, viz., Cana-
valia turgida, Vigna lutea, Cæsalpinia Bonducella, Afzelia
bijuga. Acacia Farnesiana (planted in Ceylon), Terminalia
Catappa (ditto), and Hernandia ovigera (? Malayan).
We shall therefore probably not be far wrong if we assume
that the littoral sea-borne fiora of the Chagos-Maldive-
Laccadive Archipelago is chiefly Malayan in composition,
partly derived directly by the currents in the north-east
monsoon, partly indirectly by the currents in the same mon-
soon (aided by the wind) from Ceylon or Southern India,
and partly by the currents and wind in the south-west
monsoon from the Seychelles, from which or from other
parts of the Arabian Sea the few non-Malayan forms, such
as Launæa pinnatifida, have been obtained. The main track
of the transporting agencies leads through the Southern
Maldives, which have in consequence the richest flora,
while the flora of the Laccadives is comparatively poor, in
fact poorer in these widespread forms than that of the
Malabar Coast. As this point may be raised as an objection,
let us briefly consider the Malabar flora. Thirty of the 43
enumerated species occur in it, against 22 in the Laccadives
and 29 on the Coromandel Coast. The two Malabar plants
not known to occur on the latter are Ipomœa grandiflora and
Euphorbia Atoto, both found in the Laccadives and Ceylon,
and possibly carried from the former to the Indian Coast
OF THE MALDIVE ISLANDS.
151
rather than in the reverse direction or from Ceylon. Of the
12 species n ot known from the Laccadives, 1 1 occur all round
the Indian and Malayan Coasts, and may have travelled by
land, while the remaining one, Pemphis acidula, has pro-
bably been derived from Malaya by the currents of the north-
east monsoon bringing it to the Coromandel Coast. The
Laccadives, on the other hand, have Tournefortia argentea,
Hernandia peltata, Thuarea sæmentosa, Lepturus repens,
which do not occur on the Malabar Coast, nor indeed in India
at all.
Many of the Malayan and Seychellan forms are abundant
in the Maidive Islands, but rare in Ceylon, showing that the
main track of distribution passes south of Ceylon. Such
are Suriana, Guettarda, Ochrosia, Cordia, Tournefortia,
Ipomœa denticulata, Lepturus repens.
Forms whose Mode of Introduction is uncertain.
There still remain upon the list the following forms
whose presence has to be accounted for. After each name
are placed letters indicating in order of probability the
agencies by which means it may have reached the islands.
Species probably introduced for cultivation are marked C,
those unintentionally introduced by man M, those probably
brought by sea currents S, those by birds B, those by
wind W : —
Tinospora cordifolia, M, C
Brassica juncea, C, M
Portulaca tuberosa, B, S
oleracea, B, S, M
quadrifida, M, B
Sida carpinifolia, B, M
( diffusa), B, S, M
(Urena sinuata). B, M
Hibiscus Abelmoschus, C, M
Adansonia digitata, C, S
Corchorus capsularis, C, M
(Triumfetta procumbens), M, B, S
Triphasia trifoliata, C, M, B
Zizyphus Jujuba, C, M, B
sp. ? M, C
Colubrina asiatica,' B, S
Vitis Linnæi, B, C, M
CardiospermumHelicacabum, S, M
Allophylus Cobbe, C, B
Indigofera tinctoria, C, M
Tephrosia tenuis, M, S
purpurea, M, S
(Sesbania aculeata), M, S
Desmodium umbellatum, M, S
gangeticum, M, S
triflorum, M, S, B
152
WILLIS AND GARDINER : BOTANY
Erythrina indica, C, S
(Phaseolus calcaratus), C, M
Cæsalpiîiia Bonduc ? S
Cassia Sophera, S, M
Adenanthera pavonina, C, M, S
Bryophyllum calycinum, M, C
Lumnitzera coccinea ? S
Gyrocarpus Jacquini, S, C
(Ammannia baccifera), B, M
Sonneratia acida C, S, B
Sesuvium Portulacastrum, S, B
(Dentella repens), M, B,
Oldenlandia corymbosa, M, B, S
( diffusa), M, B, S
umbellata, M, B, S
biflora, M, B, S
Morinda citrifolia, S, B
Vernonia cinerea, M, W
Adenosteiiima viscosum, M, B, S
Agératum conyzoides, M, W
(Blumea laciniata), M, W
membranacea, M, W
Eclipta alba, M, B
(Wedelia calendulacea), M, B
(Bidens pilosa), M, B
(Crépis acaulis), M, W
Tithonia diversifolia, C, M
Artemisia vulgaris, C, M
Emilia sonchifolia, M, W
Lactuca sp ? M, W
(Plumbago zeylanica), M, B
Aj-disia humilis, S, B
Chrysophyllum ? S, B
Mimusops Elengi, C, B
Vinca rosea, M, C
Calotropis gigantea, C, M
(Tylophora asthmatica), W, C
(Asclepias curassavica), M, W
(Leptadenia reticulata), W, M
Trichodesma zeylanicum, M, S
Ipomoea Turpethum, C, M. S
( sinuata), C, M
Convolvulus parviflorus, S, M
(Solanum torvum), M, B
Physalis minima, B, C, M
Datura fastuosa, M, C, B
Herpestis Monniera, B, M
(Striga lutea), B, M
Barleria Prionitis, C, M
Lippia nodiflora, S, B
Stachytarpheta indica, M, C
Premna integrifolia, B, S
Vitex Negundo, B, C, S
Ocimum basilicum, C, M
gratissimum, M,fC
sanctum, C, M
Plectranthus zeylanicus ? M, C
Boerhaavia diffusa, B, S
( repens), M, B
Mirabilis Jalapa, C, M
Pisonia aculeata, B, S
morindæfolia, S, B
Celosia argentea, M, C
Amarantus gangeticus, C, M
viridis, C, M
Nothosærua brachiata, S, M
Achyranthes aspera, M, B, S
(Polygonum barbatum), B, M
Cassytha filiformis, S, B
Euphorbia pilulifera, M, B
thymifolia, M, B
Glochidion littorale ? S, M
Trewia nudiflora ? C, M
Ficus benghalensis, B, C
infectoria, B, C
retusa, B, C
sp. ? B, G
(Casuarina equisetifolia), S, C
Tacca pinnatifida, S, C
(Dioscorea bulbifera), C, M
Pancratiuin, sp. ? S, C
Asparagus racemosus, C, B, M
Commelina, sp. ? M, C
Bhœo discolor, C, S
Cocos nucifera, S, C
Pandanus, sp. ? S, C
sp. ? S, C
Pycreus pumilus, M, S
polystachyus, S, M, B
(Cyperus compressus), M, S
( pachyrhizus), M, S
( ligularis), M, S
sp., M, S
Mariscus albescens, S, M
OP THE MALDIVE ISLANDS.
153
(Kyllinga brevifolia), M, S
( monocephala), M, S
Fimbristylis spathacea, S, M, B
( diphylla), S, B
( glomerata), S, B
Cladium jamaicense, B, S, M
Paspalum sanguinale, S, M
(Panicum trigonum), M, S
Stenotaphrum complanatum, S, M
Zoysia pungens, S, M
Ischæmum muticum, S, M
Cycas circinalis, S, C
These represent a total of 132 species, belonging to 103
genera and 45 families, of which 101 species, 82 genera, and
41 families are Maldivian. If we group them according to
the first-mentioned means of distribution in each case, we
get—
Whole List.
Maldives only.
Cultivated ?
29
26
Introduced by man ?
50
32
Introduced by sea ?
30
27
Introduced by wind ?
2
—
Introduced by birds ?
21
16
32
101
Owing to the doubtfulness in these cases as to the precise
mechanism of introduction of each species, it would be idle
to discuss the distribution as has been done above with the
species whose transport could be pretty certainly decided,
but a consideration of these species gives the same general
results as the others, and supports the general conclusions
as to distribution arrived at above, or at least does not
contradict them. Those attributed to birds and wind have
been mainly carried over short distances only, those sup-
posed sea-borne are mainly Seychellen and Malayan, the
weeds are mostly in frequented islands.
Summing up now the floras of all the groups of islands —
Chagos, Maldives, Minikoi, Laccadives — and employing the
same criteria throughout, we get the following table : —
Mode of
Transport. Chagos. Maldives. Minikoi. Laccadives. Whole Flora.
Sea currents ...
14
... 39 ...
29 ...
22
... 43)
Sea, probably...
7
... 27 ...
10 ...
11
... 30j
Birds
—
... 1
—
6
... 7(
Birds, probably
5
... 16 ...
7 ...
10
... 21 ^
Wind
7
... 4 ...
2
2
... 11)
Wind, probably
—
... —
—
2
2 3
154 WILLIS AND GARDINER : BOTANY
Mode of
Transport. Chagos. Maldives.
Minikoi.
Laccadives.
Whole Flora.
Total independ-
ent of man... 33
... 87
... 48
... 53 .
.. 114
Man, uninten-
tional ... 2
... 41
... 21
... 35 .
56 ( .Qg
Man, probably 9
... 32
... 16
... 26 .
.. 50^
—
—
—
—
—
Total wild flora 44
160
85
114
220
Cultivated ... 4
... 98
... 40
... 32 .
.. 110 1
Cultivated, pro-
bably ... 1
... 26
9
4 .
U39
.. 29j
—
■ ■ ■
_
»
—
Grrand Total ... 49
284
134
150
359
—
—
—
—
' Leaving out the
cultivated plants.
and calculating the
rest in percentages, crosswise,
, we get the following table : —
Chagos.
Maldives.
Minikoi.
Laccadives.
Sea
28*76 ...
90*41 .
.. 53*42 ..
. 45*20
Birds ...
17*85 ...
, 60*71 .
.. 25*00 ..
. 57*14
Wind ...
53*84 ...
30*77 .
.. 15*38 ..
. 30*77
Total independent of
man ...
28*95 ..
. 76*31 .
... 42*10 ..
. 46*49
Man
10*37 ..
. 68*86 .
... 34*90 ..
,. 57*54
Total wild
20*00 ..
. 72*72 .
... 38*63 ..
,. 51*81
(Thus sea-borne plants on the Chagos are 28’76 per cent,
of the whole sea-borne flora, not of the flora of the Chagos
themselves. The total wild flora of the Maldives is 72*72
per cent, of the total wild flora of all the groups.) We may
now proceed to use these figures in the same way as in
calculating insect preferences to flowers,* subtracting the
total percentage from the particular in each case. Thus, of
species unintentionally introduced by man, the Chagos,
which are furthest from the stream of traffic, have a decided
deficiency (10*37 -20 per cent.). The Laccadives have, on the
other hand, a decided preponderance of these plants (57*54-
51*81 per cent.). Dealing now with the total independent
* Cf. Willis : Manual and Dictionary of the Flowering Plants and Ferns,
I., p. 102.
OF THE MALDIVE ISLANDS.
155
of man, we find in the Chagos a great proportion of wind-
borne species (53’84-28*95)," a deficiency of bird-carried
species (17‘85-28*95), and an average of sea-borne. The Mal-
dives have a large preponderance of sea-borne species, a
deficiency of the others. Minikoi has a considerable excess
of sea-borne, a deficiency of otherwise carried forms. The
Laccadives show a slight deficiency in sea transport, and a
preponderance in bird. The total evidence thus shows as
before that the sea is the chief agent, after the unintentional
action of man, in stocking outlying oceanic islands, and the
more so the farther out they lie, or the more (as in the case
of the Maldives) they lie in the track of the great equatorial
current systems. Bird carriage is chiefly important in islands
nearest to the mainland, while wind is operative every-
where, but especially nearer to mainlands and in the carriage
of cryptogams. Unintentional introductions by man are
more numerous the nearer to the ports of commerce.
VIL--THE ORDER OF APPEARANCE AND
THE COMPETITION OF PLANTS ON
NEW ISLANDS,
Very little has been described as to the first plants to
people a new island beyond the noting of certain species
actually observed growing on newly formed beaches, e.g.,
by Darwin, Guppy, and others (see Hemsley, l.c.), Darwin
notes that Pemphis is very often among the first to appear.
Guppy notes that on the windward or growing side of a
reef there is a scanty vegetation with few trees, Casuarina,
Pandanus, &c. On the leeward, older side, where there is
a richer soil with more humus, there is a denser vegetation,
the trees forming a thick belt overhanging the rising
tide ; common among them are Barringtonia speciosa, Calo-
phyllum Inophyllum, Thespesia populnea. Hibiscus tilia-
ceus, &c. Just within the line of trees bordering the beach
* It must, however, be remembered that the equatorial climate of the
Chagos is especially favourable to ferus, which form the majority of the
plants under this head.
(21)
156
WILLIS AND GARDINER ; BOTANY
are Cycas, Pandanus spp., Heritiera littoralis, Terminalia
Catappa, Cerbera Odollam, &c. Further inland are large
banyans, and other trees introduced by birds.
Under the descriptions of the various islands and atolls
above given many interesting details are recorded as to the
composition of the floras, and the appearance of pJants on
newly formed reefs or the flora of those which are wash-
ing away. The main general features only are summarized
below. The flora differs considerably according to whether
the island in question is rocky or sandy. In each case the
more strictly littoral flora may be perhaps subdivided into
the beach jungle and the shore herbs.
When a new bank forms, the vegetation brought to it is
at first all sea-borne, but after a few shrubs appear the
islet is visited also by birds, bringing other species with
them. The order of appearance of plants was studied in
Mahlos {q.v.) on a number of islets, and confirmatory
observations were made elsewhere. The first arrivals are
in general Launæa, Spinifex, Mariscus Dregeanus, Eragros-
tis plumosa, Aerua lanata, Tournefortia argentea, Suriana,
and Scævola. The last-named is often late in coming, but
spreads with great rapidity when once established, and
largely at the expense of the Tournefortia, which, at first
numerous, is often only represented after some time by a
few clumps or isolated trees. On the first shrubs Cassytha
filiformis is nearly always to be seen, presumably brought
by birds. The islets are visited by waders, terns, herons,
&c. Thus on a young sandy islet, both a shore herb forma-
tion and a beach jungle are soon formed. The former lies
further out than the latter, and corresponds very much to
the Ipomœa Pes-capræ formation of Schimper (lx. 77).
Ipomoea biloba (I. Pes-capræ) itself is rather rare in the
Maldives. The beach jungle of the sandy island is not so
complex as that described by Schimper as his Barringtonia
formation, but resembles it in general. Ochrosia borbonica,
Terminalia catappa, Morinda citrifolia, and the small species
OF THE MALDIVB ISLANDS.
157
of Pandanus soon appear as the islet grows, and when there
is a large drier central part Guettarda speciosa and Hernan-
dia peltata are usually to be found. Pemphis also occurs,
but is only common in rocky islands. In the southern atoll
of Suvadiva the beach jungle includes much Pandanus at the
expense of the Scævola and Tournefortia. On rocky islands
the beach jungle is much loftier in growth, and includes
the large species of Pandanus, Allophylus Cobbe, Hernandia
peltata, Acalypha fallax, Premna integrifolia, Barringtonia
speciosa. Banyans, Calophyllum Inophyllum, &c.
The contrast between the marginal jungle of forming
beaches and those which are washing away (only when
latter are rocky) is very remarkable. The former have
Scævola and Tournefortia, with the shore herbs outside
them, the latter only Pemphis acidula, which may often be
found with its roots washed by the water. The presence of
this species in a continuous belt near the water regularly
indicates a washing away of a rocky beach. In the case of
a sandy islet washing away, the process is generally very
much more rapid, so fast indeed that there can be no
definite flora. Should a period of rest or of very slow
erosion set in — a period which is characterized by the
formation of large masses of sandstone in the beach —
Pemphis begins to assume the same dominance, and forms
with the sandstone an important barrier against the further
encroachment of the sea. The near approach of the sea
does not, except very slowly, kill hard-wooded trees except
banyans. Softer shrubs sicken, die, and are replaced by
Pemphis. Pemphis is rarely found until the island is
washing away. Tournefortia remains until the water
begins to actually wash its roots with every high tide, and
then it dies. Pemphis flourishes best where its roots are
regularly washed by the tide, and will stand thus for many
years. It has a remarkably hard red wood which resists
decay and all destructive organisms. Its biological position
compares with that of willows and withes grown by the
158
WILLTS AND GARDINER : BOTANY
banks of rivers to preserve them. At last when its roots
are completely submerged, as off Turadu, it too dies, but
even here it must have been in this condition for some
three or four years, so that the process is slow.
Of other trees the cocoanut lives with the waves washing
all round it at each tide until it falls. Some considerable
portion of its root mass must always be above the water.
When grown in a rocky soil it bears about three times as
well as in a sandy one, and has larger nuts. On new rocky
islands it flourishes at once, and bears in eight years, but on
sandy ones its growth is very slow and it takes about
twenty years to come into bearing. The best place for it is
the junction of the two sub-strata. Sea-borne cocoanuts
are usually bored into and destroyed by the shore and other
crabs. Pandanus does not appear readily on any rocky or
sandy bank. It is almost certainly sea-borne, and its
segments may very probably be dragged out of the reach
of the waves by the crabs. On an eroding coast as the sea
approaches root after root is killed until the whole goes.
In general it flourishes best on sand rather than rock.
It very quickly appears in clumps on old grain land, and in
places this is a perfect jungle of its stems and roots. This
again may be attributed to rats and crabs. Of large trees
the banyans do not like salt water, and as with Pandanus
root after root dies and the whole tree goes when the centre
root is reached. Bread-fruit goes still more quickly, but
Calophyllum and Barringtonia will continue to flourish
until absolutely washed by the waves. Even then their
dead trunks continue to be conspicuous features of the
beach, as for instance on Mamaduwari, S. Mahlos.
Mangroves are not well represented as a formation in the
islands. On new sandbanks or on true outer beaches they
are never found. They flourish at the heads of deep bays,
or on the inner shores of crescent-shaped islands. Where,
as in Ehasdu, Landu, and Fendikolu, Miladumadulu, they
are found round pools and lakes, it seems probable that they
OF THE MALDIVE ISLANDS.
159
owe their position to the lakes having been at some time in
direct. communication with the sea. This is probable also
on geological grounds, while the lake at Huludu, Addu Atoll,
which has no mangrove, was apparently due to other causes.
VIIL^TRE FLORAS OF OCEANIC ISLANDS.
There now remain to be discussed, very briefly, some
features of the more general bearing of the above facts on
questions of geographical distribution in general. Firstly,
as to the much discussed question of the origin of the floras
of oceanic islands. It is evident that in a tropical island,
man’s interference being supposed absent, we shall get, on
the beach, and in the case of an island with a loftier centre
also to some extent inland, the littoral sea-borne species of
the ocean in which the island lies. If not too far removed
from the mainland, bird-introduced littoral species will also
appear, and also probably some other inland species which
will become established if the island be big enough and
raised enough to permit them to grow independently of the
littorals. Wind will bring species of fern^ and other cryp-
togams, and possibly a few Compositæ or other flowering
plants. Now, suppose the island to have been in the first
place quite free of any vegetation, and consider what may
happen in the course of a long period of time. The further
history of the littorals and of the inland species will almost
certainly be quite different. The former will come in large
numbers, depending on the currents passing the island, but
the available area for their growth will be small, and the
competition greater. Continual crossing with new arrivals
will probably help to keep the type constant, and we shall
not expect to find many endemic forms developing from the
littoral flora. If any such should form and remain littoral
in their habitat, they will probably spread about the ocean
like the other species that first came to the island, and their
place of origin may be finally undiscoverable. With the
inland forms introduced by wind and birds the case will be
160
WILLIS AND GARDINER ; BOTANY
quite different. Few individuals arrive, and probably only
at long intervals of time apart, and there is no special Reason
why further arrivals should necessarily be others of the same
species. Finding a large vacant area, these forms may
spread in great abundance over the interior of the island,
and ultimately give rise to endemic species, whose chance
of further distribution is small, and depends on the nearness
of other islands. It has been elsewhere shown that wind
distribution tends to carry more species, but fewer indivi-
duals than animal distribution, and we may therefore expect
perhaps to find more wind-carried species than bird-carried
giving rise to endemics. Treub, in his work on the new flora
of Krakatoa,^ has shown that we may expect to find many
ferns in oceanic islands, due to their carriage by wind, and
supported by the actual state of the floras of Juan Fernandez,
Ascension, &c. We shall expect to find ferns and other
cryptogams specially well represented, but may also look to
find Compositæ (wind), Rubiaceæ (birds). Palms (birds),
small seeded plants (birds’ feet), and others carried in
similar ways. Examination of the actual floras of oceanic
islands will show that a large number of their endemic
species belong to orders in which means of carriage are well
marked, but, on the other hand, there are many in which it
is difficult to assume such ancestry. Thus, in the Admiralty
Islands the few endemic species belong to the genera
Medinilla, Hydnophytum, Dendrobium, Cyathea, Hymeno-
phyllum. Polypodium, Acrostichum. In Diego Garcia
occurs an endemic Asplénium, the only endemic species in
the Laccadive-Chagos chain among the higher plants, with
which alone we are at present dealing. In Rodriguez we
have a small island isolated by a considerable distance from
its nearest neighbours, the Seychelles and the Chagos. In
former times it may have formed part of a large island
lying east of the Seychelles-Mascarene Island, indicated on
the map by the 2,000-fathom line, but any land connection
with that region must have been at a vast distance of time,
* Aim. Buitenz., VII.
OF THF MALDIVB ISLANDS.
161
and would perhaps be indicated by a greater generic
endemism in the flora than is actually the case. The
endemics of Rodriguez, as we might expect from what has
been said, consist in great part of Rubiaceæ, Compositæ,
Asclepiadaceæ, Verbenaceæ, Palms, Selaginellas, &c. Some
of them show Asiatic affinities, but if we suppose land to
have formerly occupied parts of the present sites of the
Maidive and Chagos Archipelagos, there is no difficulty in
accounting for their passage. It is evident from the facts
above given that if there were but an inland part, where the
wind and bird-borne species could be free from the competi-
tion of the littorals, and from the unfavourable soil and
climate of the shore, species with any reasonably good
distribution mechanism suited to birds or wind could get
across from Ceylon to Africa or vice versâ. Taking now the
case of the Seychelles,, we find in them 60 endemics, which
include 14 Rubiaceæ, 6 Vascular Cryptogams, 6 Palms,
3 Pandani, 2 Compositæ, 2 Orchids, and others. Here,
again, the facts fit in well with the views expressed above.
There is probably no need to assume complete land connec-
tions across the Indian Ocean to explain the floras of its
islands, or perhaps even the affinities of the African and
Indian floras, but we must almost certainly assume that
formerly there were larger islands in the present places
occupied by the Maldives, Chagos, and Rodriguez (see map).
There are, as is well known, other oceanic islands, in
which, for the full explanation of the flora, it is necessary, so
far as our present state of knowledge goes, to assume former
continental connection or great land extension, but for the
islands of the tropical Indian Ocean this is perhaps un-
necessary. The present paper, however, is hardly the place
for a full discussion of the question. Let us, in conclusion,
briefly consider the bearings of the facts of the Laccadive-
Maldive-Chagos flora on the question of the former greater
land area of these groups. If we suppose that in former
times there was a large area of land, whether insular or
162
WILLIS AND GARDINER : BOTANY
continental, in place of these groups, let us consider what
would happen as it became submerged. Its flora would
inevitably get smaller and smaller as the available area and
choice of suitable habitats decreased, and presently a condi-
tion like that now seen in Rodriguez would be attained.
Let us now imagine Rodriguez to sink to submergence ; it
is all but certain that its peculiar flora would be extermi-
nated, being unable to live on the coral reef. If any species
developed into a littoral form, it would probably become
spread about the Indian Ocean, and cease to be confined to
Rodriguez or to the island in which it first appeared.
Ochrosia borbonica is probably such a case. It is therefore
evident that the flora of the Laccadives and Maldives, as we
now find it, may be equally regarded as due to the submer-
gence of the former land areas, or to the appearance of new
plants on an area appearing for the first time above the
waves, or in other words, it is valueless as evidence one way
or the other, so far as our present means of interpreting the
evidence go. Other evidence seems to render necessary the
supposition that where the coral reefs now are there was
formerly a great extension of land, but there is no evidence
from the present state of the flora for or against this view,
nor any certain evidence even of the continual presence of
land in the places occupied by the archipelagos. If these
were now to be submerged and to rise again above the
waves, they would probably acquire in time a flora almost
identical with that which they now possess. Other evidence
seems to indicate that the former land masses were complete-
ly submerged, leaving only banks covered with at least a few
fathoms of water ; the present islands subsequently rose
above the sea, to be peopled by waves, birds, wind, and
finally man, with their present flora.
IX.— GENERAL SUMMARY.
The Maidive Islands have a flora of about 284 spp. of
flowering plants and ferns (or with the Chagos and Lacca-
dives 359), of which the cultivated plants and weeds number
OF THE MALDIVE ISLANDS.
163
about 197 (245), while of the remainder about 66 (73) are
probably due to introduction by sea, 17 (28) by birds, 4 (13)
by wind.
The flora is enumerated in detail, and the local and general
distribution of its members mentioned, as well as the
Maldivian names of the plants. It contains but few plants
of special interest. Cladium jamaicense and Ardisia
humilis in the southern atolls are worthy of note.
The cultivated species in the Maldives number 100 or
more, much exceeding those in the other archipelagos. The
chief locally cultivated grain is kurakkan (Eleusine Cora-
cana), rice being imported. The most important cultivation
is that of cocoanuts. A classified description of the
economic products of the islands is given. The most
generally interesting is perhaps the mat industry, in which
Pycreus polystachyus is employed. The weeds number
about 79, mostly the common weeds of cultivation of India
and Ceylon.
Sea-borne species are especially numerous in the Maldives,
including many not found in India or Ceylon, or in the
Laccadives or Chagos. The main stream of the currents
passes through the group, and the flora is thus derived
partly from the Malay Archipelago, partly from the
Seychelles, Africa, India, and Ceylon. A full discussion of
the effects of currents and wind on the dispersal of these
plants is given.
Bird-carried species are mostly found in the Laccadives,
nearest to the mainland, but there is slight evidence to
indicate that any part of the combined archipelagos can be
reached by bird transport.
Wind is responsible for the introduction of many crypto-
gams, but there is little reliable evidence for the carriage of
other plants.
Interesting details are given of the order of appearance
of new plants on newly formed islands, and other points
of this nature.
(22)
164
BOTANY OF THE MALDIVE ISLANDS«
The bearing of the facts on general questions of the
distribution of plants and the flora of oceanic islands is
briefly touched upon, and evidence given to show the
probability that there were formerly land masses where the
coral archipelagos of the Indian Ocean now lie, though
the present composition of the floras of these archipelagos
will flt in equally well with any theory of the origin of
these reefs, which allows for the complete submergence
of the former land.
Peradeniya, October 19, 1901.
Explanation of Plate II,
Fig. 1. — Map of Western Indian Ocean, showing archipelagos, currents,
&c. Arrows indicate the direction of currents : when two occur together
facing in opposite directions, the longer indicates the current in the south-
west, the shorter that in the north-east, monsoon. The 2,000-fathom line
is marked, and indicates what may once have been the configuration of the
land masses in the Indian Ocean.
Fig. 2. — Map of the Maidive Islands. Scale, 1 inch to 60 miles.
ANN. PERAp
40"
0
ANN, PERAD. VOL, I,
PL, II,
Observations on Dracæna reflexa, Mm.
BY
HERBERT WRIGHT.
(With Plate III.)
The absence of cauline branching in the vegetative
system among the commoner monocotyledonous plants
and its prevalence among dicotyledons is a familiar fact.
In several monocotyledons, however, a definite system of
branching occurs, e.g., in the Pandanaceæ,"*^ Palmæ,t and
some tribes of Liliaceæ.
In the last-named order the occasional branching and often
bushy habit of species of Yucca and Asparagus, the more
marked semblance to an arborescent type exemplified in
species of Cordyline, and the magnificent tree-like form of
species of Dracæna are examples which may be mentioned.
The genus Dracæna has long been known as containing
species which assume an arborescent type approximating
to that of dicotyledonous trees. D. Draco, L., even when
growing under a condition of open canopy, possesses a well-
defined leader, which when old branches considerably, each
ultimate twig being terminated by a tuft of crowded linear-
lanceolate leaves. The famous specimen of Teneriffe was
one of the most gigantic, and, according to Humboldt, the
* Schumann, Die Verzweigung der Pandanaceen. Engl. Jahrb. XXIII.,
1897, p. 559.
t Drude, inEngler andPrantl, Die Nat. Pflanzenfamilien, Palmæ, p. 9.
Morris, The Phenomena concerned in the production of Forked and
Branched Palms. Journ. Linn. Soc. XXIX., 1893, p. 281.
[Annals of the Royal Botanic Gardens, Peradeniya, Vol I., Pt. II., December, 1901.]
166
WRIGHT : OBSERVATIONS ON
plant measured 70 feet in height and the stem 45 feet in
circumference. The dimensions of this unique specimen
are not to be wondered at, if we accept the opinion of
Meyen, and regard the antiquity of the tree as being greater
than that of the Pyramids. The size reached in fifty years
by the tree dealt with in this paper, however, perhaps throws
some doubt on this supposed immense age of the Teneriffe
specimen.
Dracœna refleæa, Lam.
Habit. — The specimen of D. reflexa figured at the end of
these notes was introduced into the Peradeniya Botanic
Gardens in 1847, and is therefore only some fifty-four years
old. It is a native of Madagascar, the Mascarene Islands, and
Tropical Africa, and, as far as my information goes, always
assumes an arborescent character. It is the very pronounced
branching habit of the tree which gives it its interest and
makes it worth a description and figure. No orie casually
seeing the tree at Peradeniya would suppose it to be a
monocotyledon.
At Peradeniya it is unique among the monocotyledons
in the complex nature of its branching system, and this,
together with its large proportions, renders it a conspicuous
object. It consists of a short main stem, irregular in out-
line, and measuring nearly three metres in circumference.
The stout branches given off reach a height of several
metres from the ground, and after repeated divisions th
thin twigs are terminated by crowded small leaves, usually
enclosing an inflorescence. The total height of the tree is
about ten metres, and the branches cover a total diameter
of nine metres. It has therefore the appearance of a low
branching dicotyledon, and compares in habit with such
trees as Garcinia spicata^ Hk. f.. Gardenia latifolia, Ait.,
Jacquinia aristata^ Jacq., and Gynometra cauliflora^ L. The
resemblance to the last-named is very striking, particularly
if grown under open canopy.
The other plants of D, reflexa now growing at Peradeniya
are too young to give rise to such a form, though in nearly
DRACÆNA REFLBXA.
167
every case the cauline branching is observable at varying
levels along the thin leader. As will be seen later, branch-
ing of the stem may occur in seedlings only a few months
old.
Development. — The seeds when fresh are green, circular,
and hard. If sown immediately after plucking, they ger-
minate within a period of two months, and if dried for some
time in the open before sowing, an earlier vitality is mani-
fest. The seedling is of the hypogeal type, and the cotyledon,
together with the greater part of the reserve food enclosed
by the hard testa, becomes detached at an early period. The
primary root is white and fleshy, and tapers strongly towards
the apex. It is followed at an early stage by a pair of
secondary roots, which arise from near the hypogeal coty-
ledonary node, and, though usually remaining thin and
thread like, attain a considerable length.
Subsequent to the dropping of the cotyledon, stout white
roots arise from the root stock area, and the primary root
changes to a faint red colour and gives ofiC numerous branches.
The roots formed during the flrst twelve months invari-
ably strongly assert their positive geotropism, and, with the
exception of one specimen exhibiting congenital concres-
cense, no abnormal developmental features were presented.
The aerial shoot has in twelve months’ time attained a
height of 15 to 20cm., and borne upwards of thirty leaves.
Usually it remains undivided during the first twelve
months, but a few examples were seen in which lateral shoots
developed from the lower part of the stem. This was the
earliest indication of the branching, which in older plants
gives rise to the arborescent type just referred to. It
may be worth while to mention that seeds were sown in
different parts of the Island, and the dry climate of Anuradha-
pura proved unsuitable for continued development. Even
established plants when forwarded to the branch garden in
this district failed to develop.
168
WRIGHT : OBSERVATIONS ON
The Vascular Tissues,— The genus Dracæna has been the
classic ground among monocotyledons for the study of
sliding growth and vascular cambial activity.
Respecting the cambial activity and the nature of the vas-
cular bundles produced in the roots and stems of species of
Dracæna and Yucca nothing need be said, since figures are
so frequent in almost every text book of botany. A few
points of general interest have been worked out, and may be
of supplementary value to our knowledge of this group of
plants. It will be remembered that Strasburger* worked
mainly with the stems and roots of D. reflexa^ and Scott and
Brebnerf with the same structures in D. fragrans, Gawl., and
D. angustifolia^ Roxb., together with the roots of D. Draco.
The first point of interest lay in determining the period in
the life of the plant and the particular area at which the
vascular cambium made its first appearance. The first indi-
cations of a vascular cambium were obtained in a seedling
nearly seven weeks old, and which consisted of an aerial
shoot bearing eight comparatively small leaves, and having a
basal diameter of 2 mm. ; the root system comprised a strongly
branched primary root about 70 mm. in length, together with
two young stout white roots arising from the root stock area.^
The cambium appeared in the pericycle of the very short
hypocotyl immediately below the cotyledonary node, and
from this area spread upwards into the stem and downwards
into the primary root, where the thickenings of the endo-
dermis helped to elucidate the origin layer of the cambial
tissue. In the subsequent development the vascular cambium
spreads more rapidly in the stem structures than in those of
the root. This continues at such a rate that in seedlings
seven months old the cambium has spread to a height of 6 cm.
in the stem and only just over 1 cm. in the primary root.
The products in the basal part of the stem have assumed con-
siderable proportions,”“radial rows of three or four vascular
bundles, — whereas in the root a continuous ring of cambium
* Botanisches Praktikum, 2nd ed. ; Histologische Beiträge, III,, 1891.
t The Tissues of certain Monocotyledons ; Ann. Bot. VII., 1893.
DRACÆNA REF LEX A.
169
does not exist, even near the root stock. The cambial activity
in the small part of the primary root is very irregular, and
develops apparently in response to local irritation in the
cortex, the cells of which had also begun to divide.
Starting from this stage I was able by means of repeated
peripheral measurements and microscopical examination of
the tissues, to determine the rate at which the vascular bundles
were added. In the plants examined,the cambium once started
in the stem produced a radial row of four vascular bundles,
together with the intervening parenchyma, in six months’
time. During twelve months’ cambial activity a radial
row of eleven bundles had been formed. It must be remem-
bered that since the periphery of the stem increases with
age, the same number of vascular bundles per radial row
represents an increased total activity in each successive
month.
The next point of interest was in determining the part of
the root in which vascular cambium originated.
Scott and Brebner state that in D, Draco and D. fragrans
a definite relation was found between the secondary thicken-
ing and the insertion of the branch roots. They found
stages in which there was pericyclic thickening only, limited
to the immediate neighbourhood of the rootlet insertion.
One case is noted in D. Draco^ where “ no sooner had the
secondary tissues begun to thin out in receding from a
lateral root than they began to widen again as the next
lateral root was approached.” It is further stated that “ the
chief formation of secondary tissues begins at the bases of
the rootlets and thence extends both up and down the root
and also peripherally.” It is therefore clear that for these
two species the cambium commences at the insertion of the
rootlets, and the maximum thickness of the secondary zone
is attained in this area and not at the base of the main
adventitious root itself.
This is quoted as being contrary to what is described for
D, reflexa, where, according to Strasburger, the cambium
starts and attains maximum thickness at the base of the
170
WRIGHT : OBSERVATIONS ON
adventitious root, and tapers off regularly towards the root
apex. It is further* suggested that this difference may be a
character of the species.
The almost unlimited supply of material of D. reflexa
allowed an investigation into this point. In the first place,
an examination of the roots of young plants showed that
cambial activity occurred at the base of the primary root
and existed in no other part of the root system, though a
considerable number of branches were given off. Exami-
nation of another specimen consisting of a main root 3*5
mm. in diameter and giving off a rootlet 2*0 mm. diameter
revealed a pericyclic cambium in the older part of the main
root, but which disappeared 1 cm. before rootlet insertion
was reached. From this area to the root insertion, at and
below the root insertion, no vascular cambium either
pericyclic or cortical was present. Other specimens show-
ing active cambium in all parts did not present any bulky
secondary products at the root insertion. Even if the
secondary products are present in maximum quantities at
the root insertion, does it necessarily indicate the point of
origin of the cambium from this area ? In studying the
origin of the cambium it should be remembered that there
is often a radial disposition of the cortical parenchyma,
which in the presence of cambium might give rise to
misleading notions.
Nevertheless, when one reflects over the fact that each
rootlet subsequently becomes the main root of another
system, there seems every reason to expect that the differen-
tiation of a cambium will start as in the primary and main
roots from the base, 1^., point of insertion. The cases
mentioned above can be explained on the grounds that the
cambial activity, commencing at the base of the primary or
main root, had spread so quickly that the areas of insertion
of several rootlets were affected prior to the appearance of
their own cambia. Whenever the differentiation of cam-
bium in a rootlet is on any account delayed, it is obvious
that that of the main root might easily overtake it. Further,
DRACÆNA REFLEXA.
171
the formation of the rootlet may itself be delayed until
cambial activity has spread from the base of the main root
to the area from which the rootlet subsequently emerges.
It would therefore seem most likely that in order to obtain
specimens showing the commencement of cambium at the
insertion of the rootlets material should be examined in
which the primary root is relatively young, and which,
nevertheless, bears rootlets in a comparatively advanced
condition of development.
These conditions obtained in one specimen consisting of
an adventitious young root 3*2 mm. diameter, bearing a
rootlet 50 mm. from its apex. The rootlet was compara-
tively well developed, being 60 mm. long, and therefore
extending beyond the apex of the main root on which it was
inserted. An examination of the area of insertion showed
the commencement of a pericyclic cambium on the side of
the insertion which extended to less than 1 millimetre
above the insertion and to a greater distance below, the
insertion. Here, then, was a case in which cambium
appeared independently at the insertion of the rootlet, first
being purely pericyclic, and subsequently, after a scattering
of the lignified endodermis, continued by the inner cells of
the cortex.
It would therefore appear that either of the contentions
held by Strasburger and Scott and Brebner may be valid
according to the vitality of the cambium in the main roots,
or the delayed or enhanced development of lateral rootlets
in the particular system dealt with. Though many speci-
mens were examined, the details observed were not such as
to allow one to conclude that the area of insertion of a
branch root was the centre from which the differentiation
of the cambium for the main roots proceeded. The develop-
ment of the cambium at the insertion point was more the
starting point for the rootlet itself and not for the mother
root, and in this respect would seem to differ from what
prevails in D. Draco and D, fragrans.
(23)
172 OBSERVATIONS ON DRACÆNA RBFLEXA.
One more point remains to be dealt with. According to
Strasburger the roots of D. refiexa are epinastic, the
secondary thickening beginning on the upper side and
continuing to be more vigorous there. Certainly the peri-
cyclic cambium is highly eccentric, and several areas of
activity can only be explained in terms of the idiosyncrasies
of the particular pericyclic cells. In the primary root,
which has a fairly regular and downward course, the peri-
cycle develops a cambium on one side only ; similarly, in
secondary roots and in lateral branches exhibiting positive
geotropism. The only relation that could possibly be
established was that the cambium was active on the side from
which a rootlet emerged, but as this occurred on any part of
the mother root no “epinastic ” development could be said to
exist. A similar arrangement has been described by Scott
and Brebner for D, Draco and D. fragrans, and the eccentric
development correlated with the proximity of the lateral
root.
The stele of the roots, though of the usual polyarch exarch
type, seems to be liable to variation. A cortical injury
often caused a breaking of the endodermis on one side, and
this layer then curved inwards at each end and produced an
arc similar in outline to that which monostelic systems often
assume in their transition to other forms. In cases of con-
genital concrescence a transverse section reveals two separate
monosteles, which when traced upwards into the older
portion fuse in a regular manner producing a single polyarcb
stele.
Explanation of Plate HI.
“ Dracæna reflexa, Lam.”
(Illustration reproduced from a photograph by L. Maddock, Esq., the Studio, Kandy.)
The tree is fifty-four years old ; circumference of stem = nearly H
metres ; height of tree =10 metres ; total spread of foliage = 0
metres. Scale, nat. size.
Ann. Perad. Vol. I
Pl. Ill
Apothecaries C° phot.
Sadag Sc. — Geneve
DRACÆNA REFLEXA Lam
i
i
I
\
\
I
■ j
REVIEWS.
173
REVIEWS.^
Disease in Plants.
By H. marshall WARD.
(Macmillan & Co., London, 7s. Cid.')
Thou&h the title of this book is comprehensive enough for a much
more bulky volume, yet its contents would perhaps be more correctly
summed up as “ plants in health and disease.”
Professor Ward has done much in England to advance the science of
plant pathology and to encourage others in the pursuit, and his work in
this field is of special interest in Ceylon, because his first scientific
investigation of a disease and its causations was at Peradeniya.
To the lay reader some of the book may be not easily understand-
able, but if read carefully and intelligently by agriculturists and
horticulturists, whether in temperate climates or the tropics, a great
gain must result. They will be led to a clearer understanding of
the complex problems which those who tend plants have to consider, if
they would keep their charges in as productive a state as possible.
To workers in the domain of plant pathology and therapeutics
the book will be welcome as much for the suggestions as to lines
of attack as for the record of positions already secured. It is to
be hoped that this volume, summing up as it does in a brief form the
main facts about plants in the abnormal states called disease, will
be the precursor of a work dealing with therapeutic methods and
results.
There is a mass of knowledge on therapeutic points gradually being
acquired in Germany, America, and other countries, which requires
to be put through the mill of inductive reasoning and reduced to a
series of general laws. Even in the present work a chapter might
most usefully have been devoted to some methods of investigation, and
a brief account given of one or more cases of the working out of
the life history of disease-producing organisms, by De Bary, Hartig,
or Tubeuf. Such examples of accurate tracing of causation by
* The articles which appear under this head are written primarily for
the Ceylon constituency of this Journal, and deal chiefly with advances in
Science which are of immediate local interest. . i
(24)
174
RETTF.WS.
observation and exact experiment go a long way to discourage
perfunctory explanations, which so often do duty for careful
investigation, and lead to the employment of empirical remedies.
Each page in many chapters touches on so many points on which
our knowledge is yet insufficient that we might suggest that an
interleaved copy, in which the observer either in the field or laboratory
could note down pertinent facts, should in many cases lead to the
record of many minor observations, which would otherwise be lost.
Professor Ward objects to the division of plant diseases into
“ physiological ” and “ parasitic,” since all disease is physiological in so
far as it consists in disturbance of normal physiological functions.
The separation into two groups, those induced primarily by physical
causes and those due to the effects of a parasitic organism, is sound both
systematically and for practical purposes, but “ environmental ” is
perhaps a more accurate name than physiological.
The whole question of grouping of diseases in plants is one of great
difficulty, and even the sub-grouping of the diseases due to specific
parasitic organisms is not easy. The method of arranging the diseases
by the natural order of the organisms causing them is plain sailing in
the case of diseases attributable to one chief and primary cause,
but does not help where one or more factors are correlated. There is
a chaos at present, and a system would be welcome, even if somewhat
arbitrary, which would aid the collection of data and pave the way to
a rational arrangement. In the book before us such grouping as
“ artificial wounds,” “ natural wounds,” “ excrescences,” “ exudations,”
“ monstrosities,” is open to various objections.
In the chapter on the spreading of disease nearly all agencies which
have been suspected or proved to directly distribute spores are
mentioned, and here we have a field of investigation possessing
attractive interest and of direct economic importance. If we take the
question of wind agency alone, we have a number of phenomena,
which, if properly understood, would lead us far on the road of
preventive sanitation.
Observation of meteorological conditions will bring much informa-
tion of value in forming a clear conception as to the means of
distribution of various parasitic fungi. The study of wind currents,
succeeded by moist atmospheres, flights of insects, itineraries of
animals with regard to their porterage of spores, will give results
which enable us to determine the best way of dealing with protective
wind belts of trees and selection of safe places for the cultivation of
plants specially liable to attacks of parasitic organisms.
It is not to be expected that in such a small book on so large a
subject everything can be included ; some mention, however, of the
factors in nature which make for the success of the host plant in its
fight against the parasite would have added to the completeness and
interest of the book.
REVIEWS.
175
The effect of direct sunlight — a subject on which Professor Ward has
already added to knowledge — on tissue containing bacteria or the
mycelium of fungi, the feeding of animals and insects on diseased
tissues, the production of roots from tissues not previously root forming
in cases of diseased roots, all have an important bearing on the plant
in its pathological aspect. In fact, all agencies, organic or physical,
which tend to restore the plant to the normal, are of the greatest
interest, both from a biological standpoint and that of the therapeutist.
At a time when attention is specially directed to the application of
exact knowledge to all human activities, this book will do much to
further this principle in a sphere the importance of which it is hard
to undervalue.
J. B. CARRUTHERS.
Tropical Agricultural Journals.
Journal d’ Agriculture Tropicale. Paris. Monthly. 20 francs per
annum.
Revue des Cultures Coloniales, Paris. Fortnightly. 20 francs per
annum.
Bulletin du Jardin Colonial, Paris. Bi-monthly. 20 francs pej’
annum.
Der Tropenpflanzer, Berlin. Monthly. 10 marks per annum.
Mededeelingen uit’s Lands Plantentuin. Buitenzorg, J ava. Appear-
ing at irregular intervals and prices.
Bulletin Economique de ITndo-Chine, Saigon. Monthly.
With the exception of Holland, the nations of Continental Europe
have hitherto been backward in developing tropical agricultural colonies.
This, however, is no longer the case, and the energetic efforts that are
now being put forth by France, Germany, and other nations to found
prosperous tropical colonies, and which are already meeting with a large
measure of success, deserve the careful attention of all interested in the
agriculture or the politics of this Island. Not merely are these foreign
colonies commencing to compete with Ceylon in the continental markets,
but other countries, favoured with richer soils or better climates, yet
under the English flag, are rapidly coming into prominence as producers
of what have hitherto been, or might be, Ceylon staples. This is not
the place in which to enter upon a discussion of this topic ; all that is
here intended is to call attention to some of the chief journals in which
may be found much that is of interest to the Ceylon planter, but whose
176
REVIEWS.
contents have as yet remained almost unknown to him. The numerous
English journals published in many different tropical lands contain
much that is of value and copy freely from one another, so that any
information appearing in English is soon spread over the colonies. To
judge from internal evidence, however, their editors rarely read any
article written in a foreign language, and yet a perusal of the journals
above mentioned would furnish much information that should be at the
service of English planters and others in the tropical colonies. The
economic history of the past century in Ceylon shows an alternation of
periods of great prosperity with periods of great depression, due to the
rise and fall of successive agricultural industries. It can scarcely be
doubted that this phase of enormous fluctuations is nearly over, and
that for the future we must devote our attention to industries in which
severe competition will have to be faced, and in which the victory, as
in Europe, will be to those who most intelligently apply to their work
all the resources of industry, ability, science, and politics. This has
long been recognized in the Dutch colony of Java, in which there has
been a wonderfully steady prosperity for many years, attended by less
fluctuation than in Ceylon, and in which the history of agricultural
progress is not so marked by the ruins of extinct industries and by
enormous areas of waste and almost valueless land. The finest scientific
institution in the tropics for the aid of agriculture is without doubt
that of Buitenzorg in Java, which it may be remarked is largely paid
for directly by the planters concerned, a proof that they believe the
results of scientific methods and investigations to be of direct practical
value to their industries. The results of much of the work carried on
in this institution are of value to Ceylon planters, if intelligently applied
to the different local conditions. Unfortunately the fact of their being
in the Dutch language renders them unavailable to most English people,
but the attention of the editors of English tropical journals should be
called to this mine of information. Among recent articles in the
“ Mededeelingen ” may be mentioned a detailed illustrated account of
Butin Schaap’s method of grafting coffee, which has proved of such
value in the Java industry, accounts of the common diseases of tomatoes,
coffee, &c., with methods of treatment, and many others.
The French and German colonies are endeavouring from the begin-
ning to work on scientific lines, and their journals contain much that is
of interest to us. Several excellent critical articles have appeared
dealing with the English tropical industries and calling attention to
their faults, with the view of enabling their rivals to start competition
on lines likely to be successful in the end. In the present period of
decreased profits in tea innumerable inqumes are made for “new
products.” In the strict sense such, things are now hardly to be found j
in some place or other everything has been or is being tried or is
regularly exploited. Detailed information about many such products
is to be found in the journals quoted above. All may be seen in the
REVIEWS.
177
library of the Peradeniya Gardens, in which almost every journal of
importance dealing with tropical agriculture is received. Each of the
journals mentioned above has its own particular line of work or
speciality. The Journal d’ Agriculture Tropicale is new, and aims at
giving crisp practical information, including the latest results of
scientific investigations at various institutions in all parts of the world.
The two numbers which have already appeared contain interesting
papers on Sisal Hemp and the machines used in its preparation, on the
various kinds of Castilloa(of this more hereafter), and on Camphor,
with many other things. The ‘‘ E-evue des Cultures Coloniales ” is an
older paper, and frequently contains useful articles, especially with
reference to the many plants which are staples in the older French
colonies, such as vanilla, nutmegs, &c. The “Bulletin du Jardin
Coloniale ” (a new institution, somewhat on the lines of Kew) contains
in its first number articles on the ploughs used by natives in the different
countries of the tropics, on Manila hemp, and other topics. The
German journal, “ Der Tropenflanzer,” contains many useful papers ;
among recent articles of interest may be mentioned a long paper by
Koschny on Castilloa in its native countries, dealing with the various
kinds of Castilloa, their yield, treatment, and commercial exploitation ;
a detailed and critical account of the tea industry of India by Schulte
im Hofe ; an interesting paper (the result largely of investigations
carried on at Peradeniya) by Preyer on Fermentation of Cacao, and
others. In the “ Bulletin de ITndo-Chine ” there are often papers with
an application to Ceylon, the climates and soils of the two countries
being not dissimilar.
Let it suffice to have indicated that the English journals and official
publications do not exhaust the available information of value, and that
in the severe competition in tropical industries, which is approaching,
no source of information may safely be neglected.
J. C. WILLIS.
NOTES,
179
NOTES.
Material for Demonstration of developing
Embryos (Klugia).
Better material for class demonstrations of the stages in develop-
ment of a dicotyledonous embryo than the Ceylon species of Klugia —
Notoniana and zeylanica — it would be difficult to find. By taking
seeds from capsules of all ages from those where the corolla has just
fallen to those that are all but ripe, and mounting in eau de Javelle,
the whole series of stages is easily demonstrated much more clearly
than in most of the plants used for class work. The embryo can be
seen in all stages from eight cells onwards, with its suspensor, gradually
filling the embryo sac and seed. The embryo sac, tapetum, nucellus,
and integument show with extraordinary clearness, and the gradual
destruction of the nucellus and endosperm can be easily seen. The
plant is well worth cultivation in Europe, where K. Notoniana should
succeed in wet soil, well drained, with a fairly steady temperature of
about 60° F.
J. C. WILLIS.
The New Laboratory at Hakgala Garden.
This building has been completed, and is now ready for work. It
lies in an open site on the patana south of the cultivated part of the
garden at an elevation of 5,580 feet above sea level. There is a
laboratory room facing north measuring 21 feet by 13 feet, with two
large working places, a herbarium of the hill flora, a small collection of
books, and simple essentials for botanical work. Opening from it is a
somewhat smaller living room, and beside this are two single bedrooms,
with a kitchen and room for a servant. The building is of wood, with
stone pillars raising it above the ground, and with fireplaces. It is
being furnished with all necessaries, except linen and cutlery. Visitors
must take their own servant to do the cooking and general attendance.
There is a charge of one rupee a night for the use of the building to
visitors not belonging to the Scientific Staff of the Department, who
180
NOTES,
must obtain the permission of the Director to reside there. Hakgala
occupies so admirably central a site for the study of the various types
of mountain flora of Ceylon that we may well hope to see much good
work done in this laboratory in the future. It has already been
occupied by Mr. Coomaraswamy, the Geologist, and by the Director.
Personal Notes.
The Barclay Medal of the Royal Asiatic Society of Bengal, given
annually for the best work in Biology in India or Ceylon, has this
year been awarded to Mr. E. E. Green, Entomologist of the Depart-
ment. This is the first award.
As in 1900, a large number of officials of similar departments in other
countries have officially visited Peradeniya during the year to study
the organization and working of the Department. They include Dr.
Stuhlmann, Director of Agriculture and Acting Governor of German
East Africa ; Dr. Yan Romburgh, Director of the Java Experiment
Gardens ; Mr. I. H. Burkill, Assistant Reporter on Economic Products
to the Government of India ; Dr. E. J. Butler, Cryptogàmist at the
Calcutta Botanic Gardens ; and Mr. R. H. Proudlock, Curator of
Gardens and Parks, Ootacamund, Madras Presidency,
Annals
OF THE
Royal Botanic Gardens,
Peradeniya.
FIRST SUPPLEMENT.
A HANDBOOK OF THE VEGETABLE ECOHOmC PRODBCTS
OF CEYLON, NATIVE, COLTIVATED, OR IIIPORTED.
BY
J. C. WILLIS AND HERBERT WRIGHT.
QlDlomüD :
H. 0. COTTLE, ACTING GOVERNMENT PRINTER, CEYLON.
1901 et seq.
Prefatory Note with regard to the Supplements of
the Annals of the Royal Botanic Gardens,
Peradeniya.
r
It is proposed to publish in tlie form of Supplements to
the Annals certain pieces of useful work more of the nature
of compilations than of original scientific work : for example,
the present Handbook of Economic Products, a revised List of
the Flora of Ceylon, &c. The Supplements will be included
in the subscription to the Journal, and portions of them will
appear as ready. They should be detached from the rest
before binding, as they are separately paged with a view to
being separately bound when complete. A few extra copies
of each portion are printed, and when complete will be sold
as separate works at a considerably enhanced price. Frag-
ments will not, as a rule, be sold separately, but may be
obtained by purchasing the whole number in which they
appear.
It is not intended that the number of Supplements shall
exceed the number of volumes, though it is probable that
the publication of a second supplement may commence before
the first is complete.
A Revision of the Podostemaceæ of
India and Ceylon.
family of Indian flowering plants is so imperfectly
known as the Podostemaceæ. This is somewhat sur-
prising when we remember how interesting these plants are
on account of their peculiar habitat and mode of life in the
swiftest waters of rushing mountain streams, their very
extraordinary morphological construction, resembling in
external and internal features that of Algæ or Bryophyta,
and their great variability, to say nothing of the great diffi-
culty of placing the order in its proper position in a natural
system of classification.
Living for the last six years near to the best known
locality for Podostemaceæ — Hakinda rapids on the Mahaweli-
ganga, a mile below the Peradeniya Gardens — where six
species occur in profusion, I have devoted much attention to
these plants. The interesting morphological and ecological
observations which I have made will be described separately ;
in the present paper I propose to describe the important
changes in the nomenclature and classification of the Indian
species, which I have found necessary in the course of my
work. New species have been found, old ones proved to be
identical with forms described under other names, new
genera established, and other changes made.
BY
J. C. WILLIS.
[Annals of the Royal Botanic Gardens, Peradeniya, Vol I. Pt. III., May, 1902.]
(25)
182
WILLIS : PODOSTBMACEÆ
Having by the end of 1898 brought my observations on the
Ceylon species to a state of comparative completeness, I
I
found it necessary for their complete understanding, as well
as for the sake of a full account of the order, to study the
Indian forms also in detail. With this in view, I visited in
1899 the great herbaria of Kew and Paris to examine the
type specimens there preserved. This, however, proved to be
largely time thrown away. The herbarium specimens are
almost all mere fragments gathered from the centres of dead
and shrivelled plants in the dry season of the year, and give
very erroneous ideas of the morphology of the mature living
plants. The descriptions of the species, even in the best
monographs and floras, based upon these fragments, are
marked by serious, even glaring, errors, confusions, and
omissions. It was evident that the only useful course was
to collect again for myself fresh material of the Indian
species from the rivers where they had been previously found.
I am therefore very much indebted to the Committee of the
British Association for the Advancement of Science, and to
the Government Grant Committee of the Royal Society, for
grants of £20 and £35, respectively, made to me in 1900 and
in 1901 for this purpose. In December, 1900, and January,
1901, I visited the hills of South-Western India, from the
Bombay Ghats to Travancore, and in December, 1901, the
Khasia Hills of Assam and the Sikkim Himalaya. As it
is eminently desirable that the Indian species should be
studied upon the spots where they grow at various times of
year, I have given under each an exact description of
the places in which I found it, for the guidance of future
workers.
I have to thank many friends who have helped me in my
work by collecting or supplying material. In particular Mr.,
C. A. Barber, Government Botanist, Madras Presidency, has
collected for me excellent dry and spirit material of many
species, including three new to science, in South Kanara,
Tinnevelli, and elsewhere. By the kind assistance of Prof.
Gammie of Poona I have obtained some good material from
OP INDIA AND CEYLON.
183
Atgaon, west of Poona, collected by Mr. R. K. Bhide. I have
also received dry and spirit material from Sir W. T. Thiselton-
Dyer, Director of the Royal Botanic Gardens, Kew ; from
Dr. D. Prain, Superintendent of the Calcutta Gardens ; from
Mr. T. F. Bourdillon, Conservator of Forests in Travancore ;
from Prof. Goebel of Munich ; from Miss Gulielma Lister ;
and others. To all these friends, and to Mr. R. D. Fenton of
Monica (Anamalais), Capt. D, Herbert, Deputy Commis-
sioner of the Khasia and Jaintia Hills, the Rev. P. Decoly,
Mr. F. Lewis, Mr, J. Parkin, Mr. H. F. Macmillan, and
others, I am very greatly indebted.
Type sets of material-herbarium specimens, and in some
cases also specimens dried on the rocks on which they grew,
or preserved in alcohol — are being distributed to the follow-
ing great herbaria and museums:— Kew, British Museum,
Cambridge, Edinburgh, Paris, Berlin, Munich, Vienna,
Copenhagen, St. Petersburg, Rome, Washington, Harvard,
Rio de Janeiro, Calcutta, Saharanpur, Poona, Ootacamund,
Singapore, Buitenzorg, Tokio, Durban, Sydney. A complete
illustrative set of type specimens and material (much more
complete than any of the distributed sets) is preserved at
Peradeniya.
It will be convenient to sum up here the work that has
already been published upon the Indian and Ceylon Podos-
temaceæ. The first species was discovered by Gomez in the
mountains of Sylhet (probably the Khasias), and enumerated
as No. 5,225 in Wallich’s Catalogue (1828) as Podostemon
Wallichii, R. Br. This species was again found near Cherra
Punji in the Khasia mountains by Griffith in 1835, and he
also discovered another species in a tributary of the Bogapani
in the same district—P. Griffithii, Wall. MSS. Both these
were described by him.* After an interval of ten years, my
predecessor, Gardner, in company with Wight, collected
several species in the Nilgiri mountains, to which Gardner
* Description of two species of Podostemon ... Asiatic Re-
searches, XX., 103.
184
WILLIS: PODOSTBMACBÆ
subsequently added several from Ceylon ; his paper^ contains
descriptions of Tristicha ceylanica, the first of this genus to
be discovered, and of Podostemon Griffithii, Wallichii, oliva-
ceum, griseum, subulatum, dichotomum, Wightii, rigidum,
and elongatum, together with an interesting account of the
order, some species of which he had also studied in Brazil.
The next to occupy himself with this order was Law, who
during his residence at Thana, near Bombay, made excur-
sions into the Ghats lying to the eastward. Numerous little
streams flow westward down these hills, uniting to flow
out near Salset Island. In the floras, Law’s localities are
often given as ‘‘ Salset River.” I am indebted to the
Collector of Thana for pointing out that no such river exists,
and that probably Salset Rivers ” is meant, including all
the little streams just mentioned. As much doubt has arisen
over Law’s specimens and their localities, I quote here all that
I was able to find in his correspondence with Sir William
Hooker (preserved at Kew) bearing upon this subject •
No. 300. Tanna, 30th October, 1844 : “ The genus Hydro-
bryum founded by Endlicherf on some species of Podoste^
mon described by Dr, Griffith in the Asiatic Researches is
considered by the latter botanist to be untenable, in which
Robert Brown agrees with him. The plant of which I sent
you a specimen is however a new and very distinct species
of Podostemon, which Dr. Griffith proposes to name P.
ecostatum,J the fruit not being ribbed as in the other species,
I enclose a few more specimens in fruit, and probably by
next mail may be able to send you some in different stages,
I have discovered another Podostemmea, which Dr. Griffith
considers to be the type of a new genus, to be characterized
in a paper on the Indian Podostemmeæ. At Dr, Wight’s
suggestion he proposes to do me the honour to name it after
me. Of this also I enclose a few specimens.”
* Observations on the Structure and Affinities of the Plants belong-ing-
to the Natural Order Podostemaceæ. Calc. Journ, Nat. Hist. VII., 1846.
t Endlicher, Genera Plantarnm, p. 268.
X P. Hookerianus, Benth.
OF INDIA AND CEYLON.
185
No. 301. Tanna, 30th December, 1844: ‘‘I enclose some
pretty good specimens of Lawia, as well as of a second
species which I have just discovered at a height of about
2,500 feet. There are also a few more specimens of Podoste-
mon ecostatum. Of all these I propose to send you specimens
hereafter preserved in spirits. These curious plants have
no roots, and appear to grow in the manner of Algæ.”
No. 288. Tanna, 27th August, 1847 (forwarding box of
plants) : ...... “ are in it specimens of rocks covered with the
lichen-like fronds of the two species of Podostemoneæ, of
which poor Griffith proposed to constitute a new genus, but
they evidently belong to Tristicha, and are allied to
T. ceyianica of Gardner, described in the Calcutta Journal of
Natural Plistory.”
In 1849 appeared Tulasne’s preliminary list of the order
in this he classifies the Podostemaceæ for the first time
into their groups, and describes the following Indian
species:—
Dicrœa, Du Petit-Thouars
Wallichii
dichotoma
Wightii
rigida
elongata
Podostemon, Michx.
SLibulatus, Gardn.
Hydrobryum, EndL
olivaceum
griseum
Griffithii
Mniopsis^ Mart.
Hookeriana (Podostemon
ecostatus Griff, mss.)
Laioia^ Griff, mss.
zeylanica (Tristicha
ceyianica, Gardn.)
pulchella
longipes
All these have been mentioned above, mostly as species of
Podostemon ; the last two named are those collected by
Law,
In his great monograph of the order,f published a few
years later, Tulasne describes the above species in further
Podostemacearam Synopsis Monographica, Ann, Sc. Nat., 3me Ser., t.
XL, 1849, p. 87.
t Monographia Podostemacearum, Arch, du Museum d’Hist, Nat., VI.,
1852.
186
WILLIS : PODOSTEMACBÆ
detail, with many good figures. No change is made in the
nomenclature, except the substitution of the genus Terniola
Tul. for Lawia, which had already been used as the name of
a group of Rubiaceæ, now reduced to Adenosachme.
We now come to Wight’s work,'"' some of which has already
been mentioned. In his leones he figures and describes
most of the known Indian species, of which the following
are new: —
Dicræa longifolia, Malabar, Rev. E. Johnson.
stylosa, mountain streams, Malabar, Rev. E. Johnson.
Mniopsis Johnsonii, rivers in Malabar, Rev. E. Johnson.
Dalzellia or Tulasnea (Terniola Tul.) foliosa, Lawii, pedunculosa,
all from Salset rivers, Law.
D. or T. ramosissima, Malabar near Cochin, Rev. E. Johnson.
All these and the other Indian species are figured in the
plates, but very inaccurately ; the relative sizes and shapes of
the parts are very far from right.
Beddome, in his Anamalai Plants, f figures and describes
two further species found in those mountains, Mniopsis
selaginoides and Dicræa algæformis.
Thwaites, in his Ceylon Flora, J enumerates the Ceylon
species already mentioned, and adds Podostemon Gardneri,
Harv. MS., from waterfalls at Ramboda, with a note to the
effect that this may possibly prove to be an early stage
of Hydrobryum oiivaceum, a supposition I have since found
to be correct. In the addenda to Tulasne’s monograph, a
doubtful species, Dicræa apicata, from the Nilgiris, is de-
scribed, and it is suggested that it may be the early stage of
D. rigida ; I have, however, found it to be that of Hydro-
bryum griseum, and almost identical with P. Gardner!.
Kurz,§ from material collected in Martaban by Mr,
Parish, makes a further new species, Hydrobryum lichen-
oides.
* leones Plantarum Asiat., V., p. 31, t. 1916-1920, Jan., 1852.
t Anamallay Plants, Trans. Linn, Soc., XXV., p. 223, 1862, 1865.
I Enumeratio Plantarum Zeylaniæ, p, 222, 1864.
§ Journ. As. Soc. Beng., XII,, pt. 2, p. 103, 1873 (read 5th March,
published 28th May).
OF INDIA AND CEYLON.
187
Weddell’s"*^' monograph has been the standard for subse-
quent lists and descriptions of the Indian species. He accepts
nearly all the species mentioned above. In Terniola, there-
fore, he puts seven species. In Hydrobryum, re-defining
the genus, he leaves H. Griffithii only, olivaceum and
griseum being united and transferred to Podostemon. He
puts Mniopsis selaginoides into Dicræa, makes D. rigida,
Wightii, and longifolia into varieties of dichotoma, and
establishes two new species, D. pterophylla and D.
minor, from material collected by Hooker in the Khasia
mountains. Mniopsis Hookerianus and M. J ohnsonii he puts
into Podostemon, and adds to this genus two new species,
P. acuminatus from the Khasias and P. microcarpus (Hydro-
bryum lichenoides, Kurz) from Burma.
Bentham and Hookerf follow Weddell fairly closely,
but unite Dicræa to Podostemon. Their grouping of the
order rests, like Weddell’s, on rather too variable characters,
and often breaks down in practice.
Warming, J in the second of his beautiful series of mono-
graphs of the order, describes Dicræa elongata and D.
algæformis in detail from Ceylon material. In subsequent
papers he has described Podostemon subulatus, Hydrobryum
olivaceum, P. acuminatum (this he places in a new genus,
Polypleurum, to which he also refers a new Siamese species,
P. Schmidtianum ; I do not regard this genus as tenable),
Lawia zeylanica, and L. foliosa. In the final paper of the
series he deals with the classification of the order. Podoste-
mon Hookerianus he transfers to a new genus, Griffithella,
and describes a new species, G. Willisiana, which I hardly
think is really distinct from G. Hookeriana. Lawia he
divides into three genera, Lawia (zeylanica type), Terniola
(foliosa type), and Dalzellia (ramosissima type). The
* De Candolle, Prodr. Syst. Nat. Regni Veget., XVII., p. 39, 1873
(October).
t Genera Plantarum, vol. III., p. 105, 1880.
Î Warming, Familien Podostemaceæ, I.-VI., Kgl. Dansk. Vid. Selsk.
Skr., 6 Rk. II., 1881, II., 1882, IV., 1888, VII., 1891, IX., 1899, XI., 1901.
188
WILLIS : PODOSTEMACBÆ
second of these, as we shall see, is not essentially different
from the first, but the third is a very distinct form. He
erects the very remarkable species Podostemon selaginoides,
Benth., into a new genus Willisia, re-instates Dicræa, and
retains Hydrobryum.
GoebeP' has described Lawia f oliosa from material collected
at Khandala.
Hooker,t in his Indian Flora, accepts the following
Podostemaceæ : —
Terniola zeylanica, TuL
pulchella, Tul.
Lawii, Wedd.
longipes, Tul.
pedunculosa, Wedd.
f oliosa, Wedd.
ramosissima, Wedd.
Hydrobryum Griffithii, Tul.
Podostemon dichotomus, Gare
stylosus, Benth.
elongatus, Gardn.
algæformis, Benth.
subulatus, Gardn.
Hookerianus, Wedd.
Johnsonii, Wedd.
olivaceus, Gardn.
acuminatus, Wedd.
microcarpus, Wedd.
Wallichii, E. Br.
pterophyllus, Benth.
minor, Benth.
selaginoides, Benth.
Ceylon
Konkan
Konkan
Konkan
Konkan, N. Kanara
Konkan
Malabar
Khasias
Mlgiris, &c.
Malabar, Anamalais
Ceylon
Anamalais, Ceylon
Ceylon
Konkan, N. Kanara
Malabar
Ceylon, Mlgiris
Khasias
Tenasserim
Khasias, Burma
Khasias
Khasias
Anamalais
and a doubtful species, Dicræa apicata, Tul., Nilgiris.
TrimenJ accepts the following in Ceylon, the last-named
being new to the Indian list : —
Lawia zeylanica, Tul.
Podostemon elongatus, Gardn.
algæformis, Benth. (Trim.)§
Podostemon subulatus, Gardn.
olivaceus, Gardn.
metzgerioides. Trim.
^ Pflanzenbiologische Schilderungen, I., 166., II., p. 331, 374.
t Flora of British India, V., p. 61, 1886.
j Trimen, Flora of Ceylon, III., p. 415, 1895.
§ Trimen. Journal of Botany, XXIII., 1885, p. 173.
OF INDIA AND CEYLON.
189
P. metzgerioides on examination turned out to be a very
distinct form, which I have raised to generic rank under the
name Farmeria/" in recognition of the fact that it was first
discovered by Prof. J. B. Farmer, who also in the most
generous way withdrew his intended publication of an
account of it on finding that I was monographing the
family.
Further new species have since been discovered by Mr.
C. A. Barber, viz., Podostemon Barberi, Hydrobryum sessile,
and Farmeria indica. In spite of this, however, the total
result of my work is to decrease the number of species, no
less than eight of Hooker’s list proving to be duplicates.
0?f someef the Characters used in the
Ggassitication of Podostemaceæ.
The detailed investigations of the morphology and life-
history of the Indian species, on which these remarks are
based, will be given in a separate paper. It will, however,
be convenient here to deal with the characters of these plants
in order, pointing out some of the many pitfalls into which
it is so easy to fall in dealing with these very variable and
polymorphic organisms. Many of the characters used in
floras and monographs are nearly or quite valueless for
purposes of distinguishing the forms. Instances are given
elsewhere. It is in the highest degree essential to investigate
the entire life-history and morphology of these plants before
any safe deductions as to their grouping can be drawn. My
observations suffer from this defect with respect to many
of the species, but are perhaps more complete than those of
my predecessors.
The General LifeGiistory . — This is fully described in a
subsequent paper, and, if clearly understood, will explain
much of the difficulty that necessarily accompanies any
attempt to use the dry material for classification of this order.
The seeds germinate at the beginning of the rains, and the plant
* Notes to Trimen’s Flora of Ceylon, V., p, 386, 1900.
(26)
190
WILLIS : PODOSTEMACEÆ
is submerged and purely vegetative till almost the end of the
wet season. The primary axis is commonly insignificant, and
from its base there buds out, horizontally, an organ which
may be, phylogenetically, either of “ shoot ” or “root” nature,
the thallus^ from which arise, endogenously and usually
acropetally, secondary shoots, which are at first only vegeta-
tive, but ultimately bear the flowers in those cases where
the thallus is of root nature. The morphology of the thallus
is very interesting. The secondary shoots (or branches, in
the case of “ shoot ” thalli) develop the flowers late in the
season, often changing their structure and appearance in so
doing, and when the dry weather commences the flowers
open, in some cases emerging through the water, in others
remaining closed till exposed to the air by the fall of the
water. The thallus when exposed usually becomes much
altered by withering and soon dies ; the seeds are shed on
the rocks to await the next rise of water, and the life-history
begins anew. It is thus easy to see that without study of
these plants upon the spot where they grow, and at different
times of year, the student is liable to fall into great errors
and difficulties in dealing with them.
The Primary Axis.— for the few species described
below, this is unknown in the Asiatic species and most others
of the order, the usual axes preserved in herbaria being the
secondaries. The characters of the primary axis are often
very interesting, and are probably of considerable value in
settling the affinities of the different forms, but as yet we
know too little of them to use them with much effect. It is
very remarkable that in the Asiatic forms at any rate the
primary axis hardly ever bears flowers.
The Thallus. — Most of the Podostemaceæ possess some
kind of thallus, using the term broadly to indicate the part of
the plant that creeps on the rocks or lies near to them. There
are two main types of this organ well illustrated in the Indian
forms: the thallus of “shoot” nature, seen in Lawia,^'' and
* Cf. Plates X.-XII. of subsequent paper.
OP INDIA AND CEYLON.
191
the thalliis of “ root ” nature seen in the remaining genera ;
in some the root is a mere creeping thread, in others a
creeping closely attached ribbon, or discoid body (Hydro-
bryum spp.), in others again a drifting thread or ribbon
attached to the rock only or principally at the base, or a
fucoid organ, or even a cup-like body, as in Griffithella."^"
Within the same species the form of the thallus is very
variable in detail (the polymorphism reaches an extra-
ordinary degree in GrifBthella*), but examination shows
that the main morphological features of it are very constant,
and therefore of high classificatory value, all the Dicræasf
have a more or less free drifting root thallus, exogenously
branched, with secondary shoots on the upper margins and
in the angles ; all the FarmeriasJ have closely creeping
endogenously branched thalli, with the secondary axes in
the basiscopic angles, and so on. The degree of attachment
to the rock is variable, as is well illustrated in Griffithella ;
so also is the degree of lobing or branching, as seen in species
of Dicræa and Hydrobryum, and in the shoot thalli of Lawia.
The size is also very variable. Many thalli alter very much
in form as they approach the flowering season, and still more,
as might be expected in water plants, after they have become
exposed to the air in the dry season. The herbarium speci-
mens of these plants are in very many cases merely examples
of these exposed dead thalli, and owing to the great difficulty
(practical impossibility in the case of such closely attached
thalli as those of some Lawias and Hydrobryums) of detach-
ing the thalli from the rocks are usually of the most
fragmentary description. In the vegetative season the
thallus is usually beset with secondary shoots almost to the
extreme tips, but very often many of these do not form
flowers, and when the flowers are ripening for anthesis the
thallus often breaks away , except in the florif erous part. This
is especially marked in the Dicræas, where in D, elongata the
Cf. Plates XXV., XXVI. of subsequent paper,
t Cf. Plates XVIII.-XXIV. of subsequent paper,
f Cf. Plates XXXVIII. XL. of subsequent paper.
192
WILLIS : PODOSTEMACEÆ
long leafy tips of thethalli are not fioriferons, do not become
lignified, and drop as soon as exposed, leaving the short
woody fioriferons basal portion;" in D. Wallichii,t on
the other hand, the flowers are often more distal, but the
marginal parts of the thallus are not lignified, and break
away towards the flowering season, so that the thallus
becomes linear, v/hich was before broadly ribbon-shaped,
and lobed, which was before entire. Only when the thallus
is very closely attached to the rock does it retain its original
form, and even here great changes usually occur when the
plant is exposed, owing to the fall of the leaves. There is
no Indian species which is well represented in herbaria by
specimens which clearly illustrate the morphology of the
thallus in its mature living condition.
Another complication is introduced by the great capacity
for rejuvenescence exhibited by most thalli ; when broken
or otherwise damaged, new growing points form behind the
injury and grow out in various ways.
The Secondary Shoots. — In the root thalli these are endoge-
nously formed close to the apex, in the shoot thalli of Lawia
further back. In the latter they are mere rosettes of leaves,
unless they happen to be at the extreme edge of the thallus,
when they soon form growing points just like the original
growing points of the thallus, and ultimately give rise to
flowers. In root thalli we get two types of secondary shoot,
the one illustrated by Podostemon, Willisia, and Tristicha,
the other by the remaining Indian genera. In the first, the
secondary axis grows on continuously and branches (except
in Willisia), forming new leaves, and reaching a considerable
length, finally forming several flowers (one in Willisia) in
the usual way. In the second, the axis remains obsolescent
during the vegetative period, being represented merely by a
fascicle of leaves emerging from an opening in the thallus.
Towards the flowering season the axis elongates so as to
* Cf. Plate XIX. of subsequent papei.
t Cf. Plate XXI. of subsequent paper.
OF INDIA AND CEYLON.
193
come a short distance above the thallus, and ultimately
bears one flower. At the base of the pedicel in these forms,
and at first covering and protecting it, are several bracts,
formed from the last vegetative leaves of the axis by the
enlargement of their sheathing bases and the fall of the
blades. The examination of herbarium specimens showing
these bracts in many different stages of this process has led
to many errors. Thus Podostemon acuminatus, Wedd.,
receives its characterization from having acuminate bracts,
the specimen examined having been gathered before the tips
of the leaves had fallen. Dicræa pterophylla, Wedd., is
characterized from the keel at the back of the bract ; the
species is identical with D. Wallichii, but the specimens
examined by Weddell had been taken from younger plants
in which the keel had been less disintegrated. Trimen,
in his Ceylon Flora, describes Hydrobryum olivaceum as
leafless, as in fact it is in herbarium specimens ; in reality
it is densely leafy, but the leaves fall when exposed.
The Leaves. — These are usually of the simplest descrip-
tion. They are very delicate, and in herbarium specimens are
usually missing, or have lost their tips. The interesting
change to bracts of some of the vegetative leaves at the end of
the wet season has already been mentioned. The hairiness of
the leaves in many species is noteworthy, as it is unusual in
water plants. The leaves of the Asiatic species are simple,
but in some cases, where they have been pressed tightly
together, have been mistaken for compound.
The Spathe.—Th.Q morphology of this organ is not yet quite
understood; in all probability it represents two or more united
leaves. From a taxonomic point of view the chief feature to
note is the way in which it opens to allow the escape of the
flower. In Hydrobryum and Farmeria it is usually prostrate
or nearly so, and opens chiefly on the upper side, while in the
other genera it is more or less erect, and splits at the apex
into a number of teeth. These are often so uniform as to
give the impression of a perianth, but examination of a
194
WILLIS : PODOSTEMACBÆ
number of specimens usually discloses the fact that hardly
any two are alike in this respect. In Lawia the spathe is
physiologically represented by the cupule, an organ of
apparently axial nature bearing leaves, and often closed
until exposed by the fall of the water. In Tristicha
ramosissima the construction is still more primitive and
the flowers emerge through the water like those of the
water Ranunculi.
llie Pedicel. — The young flower is enclosed within the
spathe or cupule, and has practically no pedicel. When the
spathe splits the pedicel elongates more or less, sometimes to
as much as 5-6 mm., before the flower opens, but often only
just enough to let the flower stand nearly erect in the spathe.
After anthesis the pedicel usually lengthens while the fruit
is ripening, and at the same time the outer pellucid cortical
tissue falls away or shrivels, while the central tissue
becomes woody, leaving the fruiting pedicel much thinner
than the flowering one, only the central lignifled tissue — the
vascular tissue and the inner cortex — persisting. Sometimes
part of the outer cortex remains on the lower part of the
pedicel and has been mistaken for a spathe. The length of
the ripe pedicel is variable, and the average of a large number
of specimens should always be taken ; it seems to be affected
by the rate of fall of the water, the steepness of the rock,
and other factors, and requires further investigation. The
fruits ripen very rapidly, and even after gathering may
ripen in pressing, so that the condition of the flower and
length of pedicel in herbarium material is singularly
untrustworthy. The leaves or bracts below the flower fall
away with the cortex of the stalk in many species ; this is
especially well marked in Willisia selaginoides, whose
fruits, really sessile, often appear to be provided with long*
pedicels.
The Flower. — The parts of the flower are, as a rule, very
constant in number and arrangement, but some features vary
greatly, and it so happens that considerable taxonomic stress
OF INDIA AND CEYLON.
195
has been laid on some of these variable characters. The
Tristicheæ, to which Lawia and Tristicha belong, have a
trimerons regular flower with perianth ; the depth of the
segmentation of the latter is variable. The Eupodostemeæ,
including the rest of the Asiatic forms, have a very
zygomorphic flower, dimerous, without perianth. At the
sides of the stalk of the monadelphous andrœceum there
are two small thread-like organs, variously regarded as
staminodes or as perianth ; these organs vary much in size
and length. The length of the stamens also varies, and the
number may be either one or two in the same species in
several cases. The ovary characters are fairly constant, but
the shape and size of the stigmas are very variable,
especially in Hydrobryum. In many forms the cortex of
the ovary is deciduous with that of the pedicel of the flower.
The Fruit.—TliQ characters of the fruit are of great
importance for classification in the order. The chief
distinguishing points are the presence or absence of ribs, and
the equality or inequality of the two lobes in the fruit of the
Eupodostemeæ. When the lobes are equal, the ribs on each
run down into the pedicel, and the fruit splits evenly down
the central ribs into two lobes, both of which remain on the
pedicel ; when the lobes are unequal, the ribs of one only
are decurrent into the pedicel, and the other lobe falls ofP
altogether after dehiscence. In the genera Hydrobryum
and Farmeria, the characters of the fruit, which are usually
generic, seem to be only of specific value.
The fruits ripen after gathering, so that the length of
pedicel, as already mentioned, is variable, and so also is the
size of the fruit. The ovary being usually smooth the ribs
are formed by the'Tall of the outer tissues of the wall and the
lignification of the vascular bundles and tissue near them, and
frequently (especially in specimens gathered unripe) these
tissues do not fall away properly, so that the fruit is not so
ribbed as usual, or may even be mistaken for a smooth
fruit {cf, description of Mniopsis Johnsonii in Wight, I, c,~).
196
WILLIS - PODOSTBMACBÆ
The seeds are usually very numerous, but in Farmeria we
have a genus with a very small number (two or four) of
larger seeds.
Th@ ef t§^@
Not merely is it extremely hard to place the order in its
proper position in a natural system of classification, but it is
almost equally difficult, in the present state of our knowledge,
to divide it satisfactorily into sub-orders, genera, and species.
The whole physiognomy of these plants is so different from
that of most Phanerogams, the morphology is so peculiar, the
organs are so variable, the plants are so exposed to injury by
water currents, and the condition of the vegetative plant
below water is so different from that of the dry fruiting
specimen above, that all authors who have worked only
with herbarium specimens or with small quantities of
material have fallen into errors in dealing with them. The
study of these plants is beset vdth innumerable pitfalls for
the unwary, and in no group is it so absolutely necessary to
base distinguishing characters upon the examination and
comparison of large quantities of material, and to study the
entire life-history and morphology before diagnosing the
groups. In numerous instances separate species and even
genera have been founded upon different phases in the life-
history of one plant, upon different parts of the same plant,
or upon material taken from different levels upon the rocks,
e,g.^ the genera Lawia and Terniola above-mentioned are
separated upon the morphology of their thalli, one being
interpreted as of root, the other as of shoot, nature. In
reality the plants examined are only varieties of one species,
agreeing in all essential details, but in the one case the
centre, in the other the growing apex, of the thallus was
examined. The cases of Dicræa pterophylla (= Wallichii)
and Podostemon acuminatum have been already mentioned.
The genus Hydrobryum as re-defined by Weddell is based
on the fan-like stigma of the species H. Griffithii, the only
species left in the genus by that author ; but if a number of
OF INDIA AND CEYLON.
197
specimens of this plant be examined, it will be found that
all kinds of stigmas occur, down to the almost subulate ones
that occur in other species, such asH. lichenoides. Hydro-
bryum olivaceum. Tub, has a fiat lichen-like thallus when
mature, but when young has an erect primary axis ; the
latter has been described as Podostemon Gardner!, Harv.,
and again as Dicræa apicata, Tal. Other instances could be
quoted, but these will suffice. The primary di vision of the
order into sub-orders and tribes is the easiest. Omitting the
Hydrostachyaceæ as forming a separate order, the family
separates naturally into two main groups, the Chlamydeæ
and Achlamydeæ of Tulasne, distinguished mainly by the
presence or absence of perianth. The former may be again
separated into two small groups, the trimerous Tristicheæ
and the pentamerous Weddellineæ. The latter may be
arranged in a series of groups of increasing zygomorphism
of flowers. Tulasne simply divides them according to the
fruits into Isolobeæ or Eulacideæ and Anisolobeæ or
Podostemoneæ. Weddell divides them into Mourereæ (sta-
men whorl complete) and Neolacideæ (incomplete whorl), and
Bentham and Hooker adopt practically the same grouping,
adding Weddellina to the Mourereæ. Warming in the
“Naturl. Pflanzenfamilien ” makes three groups, Mara-
threæ, Mourereæ, Eupodostemeæ, the last being the
most zygomorphic in the flowers, but including isolobous
and anisolobous fruits, and being further divided on this
character into two smaller groups. In his recent paper
already quoted, he divides the group into Apinagieæ,
Mourereæ, Marathreæ, and Eupodostemeæ or Podostemeæ,
sections of gradually increasing zygomorphism. We may, I
think, regard this as the nearest approach at present possible
to a natural grouping, but we must acquire a much more
detailed knowledge of the order before we can draw up a
good classification.
All the Asiatic species known belong to the Tristicheæ
and Eupodostemeæ, and thus split into two widely separated
groups.
(27)
198
WILLIS : PODOSTEMACBÆ
The Asiatic Tristicheœ. — All the Asiatic species are
referred by Weddell and Hooker to the genus Lawia
(Terniola), but the morphology of the peculiar South Indian
species L. ramosissima, Wmg.,is so essentially different from
that of the Lawias of the L. zeylanica type, that I prefer
to place it in a separate genus, following Warming ; I do
not, however, think that it needs a genus to itself as
suggested by him. It may very well form a sub-section of
the almost cosmopolitan tropical genus Tristicha, and
probably represents a more primitive form with relation-
ships on one side to the other species of Tristicha and on
another to Lawia. In many ways this plant may be looked
upon as perhaps the most primitive in organization of all
the Podostemaceæ, and in habit and mode of life it comes
more near to the common type of water plants than any of
the rest of the order. Including this plant in Tristicha of
course involves a slight revision of the generic characters,
and it may perhaps be well also to split the genus into two
sub-genera, Dalzellia and Eutristicha.
The Asiatic Eupodostemece (in Warming'' s sense). — All the
other Asiatic species belong to this group. The group itself,
as indicated above, is a fairly natural one, but its further
division into genera and species is very difficult, and will
only be practicable with any degree of accuracy when we
shall have acquired a much more detailed knowledge of the
structure and life-history of its members. Warming in
1890^" divided it according to the fruit into Isolobæ {Bicrœa^
Du Pet. Th., Hydrohryum, Endl., Geratolacis^ Wedd., Ango-
Icea, Wedd.) and Heterolobæ (Mniopsis^ Mart, et Zucc.,
Podostemon^ Michx., Oserya^ Tul. et Wedd., Castelnavia^Tvd.
et Wedd., SphcerothylaXy Bischoff, to which have since been
added Leiothylax^ Wmg., Gladopus, Hj.. Moll., Farmeria^
Willis, Willisia, Wmg., Griffithella^ Wmg., Polypleurum^
Wmg.). Before going on to the rest of the taxonomy, we
must deal with these genera, and consider their right to
autonomy. Angolæa is an apparently well-marked African
* Engler and Prantl, Die Natürlichen Pfianzenfamilien.
OF INDIA AND CEYLON.
199
genus, without representatives in Asia and America.
Sphærothylax is a characteristic Abyssinian genus, and
Leiothy lax is near to it. Cladopus, another very peculiar form
in many respects, appears to be a good genus, confined, so
far as we yet know, to Java. Castelnavia is probably a
compound or polyphyletic genus, but at least has no repre-
sentatives in Asia. We must consider the others in more
detail.
All the American Podostemons, so far as we know them
(chiefiy from the beautiful work of Professor Warming)
have long cylindrical or slightly flattened narrow creeping
roots, bearing complexly-branched many-flowered secondary
shoots, whose leaves are often compound, and which show
no formation of scales or bracts in the way seen in most of
the forms found in India. The capsule is ribbed, with one
deciduous valve (see figures in Warming, lx. I., III., V.). The
x^merican Mniopses have similar morphological construction,
but smooth capsules and multifid stigmas with large
papillæ {lx. I., III., V.). It seems therefore that the latter is
a genus of which there are no Asiatic representatives, unless
Podostemon Hookerianus, Benth. (Mniopsis Hookeriana,
Tul.), or P. selaginoides, Benth. (M. selaginoides, Bedd.),
belong to it. Neither of these has the stigma of the
American forms, nor the vegetative morphological construc-
tion, and both come as near to other genera as to Mniopsis
proper. I think, therefore, that we may provisionally accept
Mniopsis as a good genus, characterized especially by the
capsules and stigmas, and confined to America. Oserya §
Devillea, also American, comes near to it, but has not the
same stigma, and has only one stamen. Probably it is
distinct, but further study is needed, and it does not con-
cern us here ; Oserya § Euoserya is evidently a Podostemon,
differing from the other species only in being monandrous.
Turning to Podostemon, we soon find that the only Asiatic
forms which approach the American in all the characters
mentioned above are Podostemon subulatus, Gardn.. in
Ceylon, and P. Barberi, Willis, lately discovered in south-west
200 WILLIS : PODOSTEMACBÆ
India. Later research may exclude these, but for the
present they may best be regarded as species of Podostemon.
They have simple leaves, but some of the American species
have the same, and they show many points which, as
compared with most of the Indian forms, seem primitive ;
perhaps their position in the extreme south-west also
indicates their antiquity.
Under Podostemon, Warming and other authors include
the curious flat lichen-like thalloid species, olivaceus, John-
sonii, lichenoides, and the new species sessile. I have
shown below that these forms are perhaps better placed
under Hydrobryum, along with the solitary species H.
Griffithii, left in it by Weddell, Hooker, Warming, &c., as the
stigmatic character of this species is not constant and cannot
be used as generic ; this genus will thus include species
with smooth and ribbed fruits, and with iso- and hetero-
lobous fruits. In most cases these difiterences are almost
generic, but in this case they seem of less importance. A
new genus, coming near to the thus extended Hydrobryum,
is Farmeria, Willis, with one species in Ceylon (with a
sessile, indéhiscent smooth fruit) and one in South India
(with a stalked, ten-ribbed, dehiscent, anisolobous fruit).
This is well marked, so far as we yet know, by its thallus
morphology and few seeds.
Now, take the genus Dicræa, merged in Podostemon by
Bentham and Hooker ; is it to stand, and if so, what is it to
include As defined by Tulasne, it has a ribbed isolobous
fruit with both valves persistent, with one-flowered secondary
shoots on a more or less dimorphic ribbon-like root
thallus. Apart from the fruit, the flower resembles that of
Podostemon. Under this definition comes a large group
of nearly allied Indian species, fairly sharply marked off
from any others : D. elongata, Wallichii, minor, dichotoma,
stylosa. All these are very variable plants, but agree
closely in many characters, possessing an exogenously
branched ribbon-like thallus, partly floating freely in the
water, with small one-flowered secondary shoots, leafy at
OF INDIA AND CEYLON.
201
first, afterwards with scaly bracts formed by a meta-
morphosis of the leaf -bases, and equal-valved 8- (or 9- to 12-)
ribbed fruit, both of whose lobes persist. Taking all
characters together, the Dicræas are a very distinct group,
probably, from their isolobous fruit, as old at least as the
Podostemons proper, and I am inclined to retain the genus,
not to merge it in Podostemon. Possibly P. Barberi comes
near to Dicræa, but it has an unequal-valved fruit and no
bracts, so that the resemblance is not so close as at first
thought appears. The South American genus Ceratolacis
comes very near to Dicræa, and perhaps should be united
with it, but detailed study of the former is required.
There remain now the two very peculiar species, selagi-
noides and Hookerianus, placed under Podostemon in
Hooker’s Flora, under Dicræa and Podostemon by Weddell.
If either is to be placed in the former genus, its special
characters require complete revision, because both have
smooth fruits with one deciduous valve. Their characters
differ in many other points from those of the Dicræas, and
there seems to be no reason for placing them in that genus.
Here, as so often in dealing with the Asiatic species,
Weddell’s work bears marks of haste and want of due
consideration, so necessary in a puzzling group like this.
Bentham and Hooker place both of these species in Podoste-
mon, apparently simply because Weddell placed them in
Dicræa and Podostemon, genera which they unite. They
differ as much from the rest of the Podostemons as from the
Dicræas. Warming gets over the difficulty by giving each
a separate genus (Griffithella Hookeriana, Willisia selagi-
noides), and they are so distinct from most other forms of
the family that this course is perhaps the best until we have
a fuller knowledge of the group. They show relationships
to Podostemon, Mniopsis, and perhaps other genera, and are
probably ancient survivals.
The genus Polypleurum, Wmg., is, I think, untenable,
resting on insufficient knowledge of the morphology of the
species proposed to be included in it. The position of the
202
WILLIS : PODOSTBMACEÆ
Siamese species P. Schmidtianum, Wmg., cannot be settled
till we know its fruit : P. acuminatum, Wmg., is a Hydro-
bryum (lichenoides, Kurz), and P. Wallichii, Wmg., is a
Dicræa.
We thus get the following as a preliminary grouping of
the Eupodostemeæ : —
Asiatic Species.
Dicræa, Du Pet. Th. elongata, dichotoma, minor, Wallichii,
stylosa.
Ceratolacis, Wedd.
Podostemon, Michx. subulatus, Barberi, (?) Schmidtianum.
Mniopsis, Mart, et Zucc.
Oserya, Tul. et Wedd.
Castelnavia, Tul. et Wedd.
Griffithella, Warming. Hookeriana.
Willisia, Warming. selaginoides.
Sphærothylax, Biscliojf.
Leiothylax, Warming.
Cladopus, Hj. Moll. Nymani.
Hydrobryum, Endl. Griffithii, sessile, olivaceum, Johnsonii,
lichenoides.
Farmeria, Willis. metzgerioides, indica.
The Generic and Specific Limits of the Asiatic
Forms.
In an order of plants of such unusual features, such
variability, and such polymorphism, it is of course impos-
sible to draw the specific limits with any accuracy or certainty
until the life-histories and the morphology of the various
forms have been studied in detail. There is little doubt in
my mind that the number of distinct forms will ultimately
prove to be very great, and that there will be much dispute
over their claims to specific rank, as with many other
variable forms. Weddell* and Goebelf have already
called attention to the fact that in the large rivers of
South America the forms are different at each successive
* Sur les Podosfcemacées en général et leur distribution géographique en
particulier. Bull. Soc. Bot. France, XIX., 1872, p. 50.
f Pflanzenbiologische Schilderungen, II., p. 333.
OF INDIA AND CEYLON.
203
cataract, perhaps specifically. My own work on the Asiatic
forms leads me to believe that this is generally true, and
that when we shall ultimately have acquired a detailed
knowledge of their morphology and life-history, we shall
find that each river or group of rivers, if not each branch or
section of each large river, has its own peculiar forms of
each genus or species represented in it. Even from the
comparatively small amount of material in my hands, a
great number of species might easily be made by admitting
as specific the well-defined differences between them. A
vast amount of comparative work on the spot is required to
thoroughly sift the question, and I therefore refrain from
complicating the already tangled mass of taxonomic litera-
ture by making numerous new species. I have rather,
probably, erred in the opposite direction by drawing my
specific limits very widely indeed. Under each I have
mentioned most of the well-marked forms observed by
myself in the various localities visited.
With regard to the genera, I have adopted a somewhat
different course, and have perhaps drawn my generic limits
too narrowly, my object having been to indicate the groups
into which the species naturally divide themselves. Probably
further study and investigation of the doubtless numer-
ous undiscovered species of South-east Asia and tropical
Africa will bridge over many of the gaps, and clearly show
which genera are to be finally retained as natural.
Podostemaceæ Indices.
Flowers hermaphrodite, regular 3-merous with perianth, or
zygomorphic 2-merous without perianth, hypogynous, small,
inconspicuous, anemophilous. Perianth when present (3),
imbricate, sepaloid, equalling or exceeding ovary, marcescent.
Achlamydeous flowers included before opening in a closed
spathe, which opens irregularly at the tip or by a slit on the
upper side. Stamens hypogynous, in regular flowers 3
204
WILLIS : PODOSTBMACBÆ
alternate with perianth segments, in irregular flowers 1-3,
usually 2, monadelphous on the lower side of the flower,
the common stalk usually much exceeding the partial
fllaments. Staminodes or abortive perianth segments in
zygomorphic flowers usually 2, filamentous, on either side
of andrœceum, with sometimes a third between the partial
filaments. Anthers introrse, 2-locular. Pollen grains
didymous in zygomorphic flowers. Ovary superior, usually
more or less ellipsoid, often oblique, 3- or 2-locular with
thick central placenta and very delicate septa ; the outer wall
usually smooth or only very faintly ribbed, often with deci-
duous cortex. Ovules anatropous, usually oo , and very minute,
few and larger in Farmeria. Stigmas as many as carpels,
sessile, subulate, dentate, or cuneate, often lobed. Capsule
usually pedicellate, the pedicel of the flower lengthening
after fertilization and at the same time becoming woody
and losing its deciduous pellucid cortical tissue, ribbed or
smooth, septifragal (closed in one sp. of Farmeria), with
equal or unequal lobes, in the former case both, in the latter
the larger, persistent on the pedicel after dehiscence. Seeds
usually ao , small, with mucilaginous outer coat. Endos-
perm 0. Embryo straight, simple.
Herbs of rapids and waterfalls in mountains of India,
Burma, and Ceylon, usually annual, submerged, and closely
attached to the rocks, flowering when exposed in the dry
season by the fall of the water level, with the habit and
structure of algæ, lichens, mosses, and liverworts. Germi-
nation with the onset of the rains. Primary axis various,
early giving rise to a thallus, in Lawia composed of flattened
shoot, in the other genera of a root structure which assumes
various forms, thread-like, ribbon-like, disc-like, cup-like,
&c., attached to the rock at all points or only at the centre,
or at the centre and two or three outer points, bearing
the leafy endogenous secondary shoots on the margins or on
the upper surfaces. Secondary shoots leafy only in the
vegetative season ; leaves usually more or less distichous,
except in the Chlamydatæ, simple, usually sheathing, often
OF INDIA AND CEYLON.
205
hairy on the upper surface ; axis in most forms elongating
only towards the flowering season, when the leaves usually
enlarge at the bases to form, bracts, whose leafy tips fall off.
Flowers usually terminal, on the thallus apices in Lawia, on
the secondary shoots in the other genera, enclosed in
spathes in the achlamydeous forms, one or more on each
shoot, sessile or pedicellate, usually not opening till exposed
to the air.
FI. regular. Perianth (3). Stamens 3.
G (3) ... ... GHLAMYDATÆ.
zygomorphic, naked. Stamens (1-
3). G (2) ... ... ACHLAMYDATÆ.
GHLAMYDATÆ-^
Thallus a thread-like creeping
root ; secondary shoots long,
complexly branched, freely
floating, bearing ramuli or
moss-like shoots of limited
growth. Floral axis sub-
tended by few leaves, some-
times slightly connate ... Tristicha.
Thallus frondose, foliiferous,
creeping, closely attached to
rocks ; secondary shoots
of small rosettes of leaves
on upper side of thallus.
Floral shoots arising from
axial cupules ... Lawia.
AGHLAMYDATÆ (all belong to
Eupodostemeæ) —
Fruit ribbed, iso- or aniso-lobous,
dehiscent, with co seeds. Spathe
erect, more or less cylindrical,
opening at apex by several teeth.
(28)
206
W J h r^IS : P( )D< >ST K M A C P Æ
Thallus fiicoid or dimorphic,
usually freely floating ;
secondary shoots 1 -flowered
with scaly bracts ; fruit
isolobous
Thallus thread- or ribbon-like,
creeping ; secondary shoots
erect, with several flowers ;
bracts dithecous, not scaly ;
fruit anisolobous
Fruit smooth, anisolobous, dehis-
cent, with oo seeds. Spathe erect,
toothed or bifid.
Thallus fucoid or attached and
closely creeping ; secondary
shoots small with distichous
leaves
Thallus crustaceous, attached to
rock, secondary shoots large,
erect, with 4-ranked leaves.
Spathe bi-lobed at tip
Fruit ribbed iso- or aniso-lobous
dehiscent, or smooth anisolobous
dehiscent or indéhiscent, with oc
or few seeds. Spathe more or less
prostrate, splitting along the upper
side. Thallus closely attached to
rock.
Thallus crustaceous or ribbon-
like, exogenously lobed or
branched ; secondary shoots
usually prostrate when flor-
iferous, 1-flowered with small
scaly bracts. Seeds oc , fruit
dehiscent, iso- or aniso-lobous,
ribbed, or smooth
Dicrœa.
Podmtemon.
Griffithella
WilUsia.
Hydrohryum.
OF INDIA AND CEYLON.
207
Thallus ribbon-like, endo-
genously branched ; secondary
shoots as in Hydrobryum,
but behind the branches of
thallus. Seeds 2-4, fruit
dehiscent or not, ribbed or
not ... ... Farmeria.
tristicheæ:.
TRISTICHA, Du Pet. Th. Nov. Gen. Madag.,p. 2, No. 8,
1806; Endl., Gen. Plant., p. 270, No. 1835; Presl.. Rel.
Haenk., p. 86 ; Meisner, Vas. Gen., p. 122 ; Gardn., Calc. J.
N. H., VII. , p. 177 p.p. ; Tul. in Ann. Sc. Nat., ser. 3, t. XI.,
1849, p. Ill, and Monographia Podostemacearum, 1852?
p. 179 ; Dufour ; Dufourea^ Bory in Willd. Sp. PL, pt. 5,
p. 55, Willdenow, A. St. Hil., A. Rich., Spreng., Roem.
et Sch,, non H. B. K., nec Delise ; Philocrena^ Bong. ;
Dalzellia, Wight, p.p.. Warming ; Terniola (Tul.), Wedd.,
p.p. ; Lawia (Tub), Wmg., 0. Kuntze, p.p. ; Tulasnea,
Wight, p.p.
Perianth (3). Stamens 1 or 3, alternating with peri.inth
segments. Carpels (3), stigmas 3. Roots creeping ; shoots
not thalloid, much branched, many-flowered.
Perianth 3-partite hypogynous, imbricate, sepaloid, membranous,
marcescent. Stamens 1 or 3, alternate with perianth segments ;
filaments slender ; anther ovate, pollen globose. Ovary superior, free,
ellipsoidal, 3-gonous, 3-locular, with thick placentæ and go anatropous
ovules ; septa thin and fragile. Stigmas 3, linear or subulate, often
with long papillæ or hairs, marcescent. Capsule triangular, 9-ribbed,
3-locular, septifragal, equi-valved. Seeds go .
Submerged herbs with the habit of Fontinalis and other aquatic
mosses. Primary axis unknown ? Roots creeping, filamentous, attached
by feet or haptera. Secondary shoots go , endogenous in the roots,
often more or less paired on opposite sides, floating freely in the
water, when developed in their highest form often long, many-flowered,
and frequently branched, but often quite short, unbranched or nearly
so, and 1-flowered or vegetative only. Branches of two kinds, long
branches repeating the structui-e of the main axis, and short branches
208
WILLIS : PODOSTEMÂCBÆ
or ramuli, consisting of a delicate axis with qo small leaves, often
3-stichous. Flowers terminal on pedicels subtended by 2-3 large
or several ordinary leaves, in the latter case the upper ones more
or less united. Pedicel and ovary with deciduous cortex, which
shrivels after flowering. Flowers small, usually (?) emerging through
the water at the beginning of dry season.
Cosmopolitan tropical. South America, Mexico, Africa.
Madagascar, India.
§1. Dalzellia, { Wight) (as genus). Stamens 3.
Leaves of ramuli not tristichous. Flower subtended by
several ordinary leaves, the upper united. India.
1. T. ramosissima, ( Wight) Willis,
§2. Eutristicha, Willis. Stamen 1. Leaves of ramuli
more or less tristichous. Flower subtended by two or three
larger leaves. America, Africa.
2. T. hypnoides, Spr.
3. T. alternifolia, TuL
TRISTIGHA RAMOSISSIMA, (Wight) Willis ; Lawia
ramosissima, Warming, in Engl. Prtl. Nat. Pflanzenfam.,
O. Kuntze ; Terniola ramosissima, Wedd., in DC. Prodr.
Dalzellia ramosissima, Wight, Ic. PL As., p. 35, 1852 ;
Tulasnea ramosissima Wight, 1. c., t. 1,920.
Stamens 3. Ramuli not tristichous. Upper leaves of floral
shoot connate. Stigmas 3, long, filamentous, hairy.
Primary axis unknown (?). Roots cylindrical, filamentous, creeping,
frequently branched, and attached to the rock by root hairs, haptera, or
disc-like feet. Secondary shoots go , endogenous at the proximal ends
of feet or haptera, floating freely in the water ; when fully developed
large — to 25 cm. long— and frequently branched, but often represented
only by a ramulus or a tuft of ramuli, or by a short 1 -flowered shoot.
Leaves at apex of secondary axis small, narrow, simple, thin, in many
ranks. Branches of two kinds, short branches or ramuli, developed
first in the leaf axils, consisting of slender filiform axis of limited
growth with co entire simple thin linear or filamentous leaves irregu-
larly arranged, and long branches, developed later above the ramuli and
OF INDIA AND CEYLON.
209
repeating the structure of the main axis. Flowers usually terminal
on short shoots bearing a few leaves, the upper usually more or less
connate, and two or more of the lower with ramuli in them axils, erect,
emerging through the water, anemophilous. Pedicel 3-5 mm. lengthen-
ing to 5-15 mm. in fruit. Perianth sepaloid, marcescent. Stamens 3,
much exserted, on flexible filaments. Anthers ovate-oblong, bilobed.
Ovary superior, at first much shorter than perianth, but after
fertilization elongating to the same length as the perianth, slightly
trigonous. Stigmas 3, long, hairy, the tips usually deciduous. Capsule
9-ribbed.
In streams of the Western Ghats 300-4,000 feet, often in
gently moving water, rather common. First discovered by
the Rev. E. Johnson. S. Kanara, at Beltangadi, 0. A. Barber,
S. Ind. Flora, No. 2,517, 2,518 ! Malabar, in streams near
Cochin, Rev. E. Johnson ! Anamalai Mountains, in the
Sholai Aar, Monica Estate, at 4,000 feet, J. C. Willis ! Travan-
core, at Mundakayam, 300 feet, T. F. Bourdillon !
All these forms show slight differences in size and shape
of leaves, number of leaves on the ramulus, and other
features, but much further study of material from many
localities is needed before the varietal rank of the different
forms can be decided.
LA WIA (Griff. MS.), Tul. in Ann. Sc. Nat., 3me. ser., t. XI.
p. 112; Warming, in Engl. Pr. Nat. Pfianzenfam. ; Trimen,
FI. Ceylon; non Wight (=Adenosachme); Terniola^l^ul.
Monogr. Pod., p. 189 ; Hooker, FI. Br. Ind. ; Wedd. in DC.
Prodr. ; Tulasnea, Wight in Ic. PI. As., t. 1,919 ; DalzeUia,
Wight, I.C., p. 34; Thwaites in Enum. PI. Zeyl., p. 223;
Tristichce sp., Gardner in Calc. J. N. H., VIL, p. 177 ; Mnian-
thus, Walp. Ann. Bot. Syst. III., 443.
Species ramosissima excluded. Specific names unaltered
under the various generic names.
Perianth (3). Stamens 3, alternating with the perianth
segments. Carpels (3). Stigmas 3. No roots. Thallus
210
WILLIS : PODOSTEMICEÆ
frondose, branched, with rosettes of leaves endogenous on
upper surface. Flowers c>o at apices of lobes, emerging
from axial cupules covered with small leaves. Fruit
9-ribbed.
Perianth 3-fid, regular, imbricate, membranous, brownish-coloured,
marcescent. Stamens 3, hypogynous, alternating with the segments of
the perianth ; anthers 2-lobed, intforse, opening by longitudinal
fissures ; filaments oval in section. Pollen globose. Ovary of 3 carpels
syncarpous, superior, ellipsoid, obtuse, more or less 3-angled, 3-
locular, with oo anatropous ovules on a thick axile placenta ; septa very
thin. Stigmas 3, more or less linear, with large papillae, marcescent,
simple. Capsule more or less enclosed in the persistent perianth, 1-3
mm. long, ellipsoid to obovoid, 9-ribbed, 3-locular, septifragal ;
valves 3, equal, opposite to the perianth, segments incurved after
dehiscence, all ultimately deciduous. Seeds very numerous, minute,
exalbuminous ; outer layer of testa becoming sticky when wetted.
Embryo straight, with crumpled cotyledons.
Herbs with frondose thalli, living on smooth rocks in rapids and
waterfalls of the mountains of Western India and Ceylon, submerged
in the period beginning with the onset of the south-west monsoon, and
flowering in the dry season when exposed by the fall of the water
Thallus closely attached to the rock, sub-orbicular or stellate, of shoot
nature, rootless, attached by root hairs at all parts, usually much
branched, with ribbon-like or fiabelliform lobes ; marginal growing
points prostrate or slightly ascending. Leaves very numerous, simple,
entire, usually acute, green or red with white-looking mesial line, very
thin and delicate, and easily detached or broken in herbarium speci-
mens : occurring in two positions, at or near the growing apices
on the upper surface and edges of the thallus, and in closely packed
endogenous rosettes on the upper surface of the older parts of the
thallus. Leaves dimorphic at the growing points, the lateral ones
broader and distichous, with the posterior margin uppermost, the
upper ones linear and irregularly arranged ; size variable, from 1-16
mm. long and to 3 mm. broad in the largest laterals. Older parts of
the thallus leafless, except for the endogenous rosettes, but marked
with scars of leaves. Flowers terminal on the growing points of the
margins of the thallus, rarely on the upper surface, solitary, on pedicels
emerging from terminal axial leafy cupules. Flowers radial in struc-
ture, but cupules thicker and shorter on the upper side, more or less
ascending, bristly in herbarium specimens. Floral axis sometimes
almost or completely foliose. Leaves on lower side of cupule larger,
distichous, as at vegetative apices. Pedicel at anthesis about 3-6 mm.
long, covered, as also is the ovary, with a pellucid cellular cortex,
which falls off after flowering, the inner tissues becoming lignified and
forming an elastic pedicel to the fruit, 5-25 mm. long, erect.
OF INDIA AND OEYLON.
21]
Seedling at first minute, tuberous by the swelling of the hypocotyl,
with flat green membranous cotyledons, without erect primary axis,
the growing point immediately growing out laterally to form the thallus
by continual branching.
It is impossible to monograph this genus from the existing
herbarium material ; the thallus is so closely attached to the
rock that it cannot be removed entire, and the plants are very
variable in their morphological features. Seven species are
recognized by Weddell, and by Hooker, who follows him, but
any one who has ever tried to identify a form from the
descriptions must have suspected the genuineness of the
species. I have transferred to Tristicha, above, the most
aberrant form, L. ramosissima, placed under this genus by
the authors mentioned, and by careful examination of
living, alcohol, and herbarium material, have reduced the
other six species to one. These six are : —
L. zeylanica, Tul.
pulchella, Tul.
longipes, Tul.
L. Lawii (Wight), Wmg.
peduiiculosa (Wight), Wmg.
foliosa (Wight), Wmg,
L. zeylanica was discovered and described by Gardner in
Ceylon. The other five are all based on Law’s material,
collected in the Ghats, east of Bombay, as mentioned above.
In Law’s letters, above quoted, Ave have seen that he him-
self only collected two forms, so far as the letters show. He
evidently sent material of these to Wight, and also to Sir
W. J. Hooker, from whom Tulasne obtained them for his
monograph. Both Tulasne and Wight have left type packets,
which I have examined at Kew and Paris. All of these
contain labels in Law’s handwriting, either “ Lawia No. 1 ” or
“ Lawia No. 2.” No other numbers occur. Tulasne named
these L. longipes and L. pulchella respectively ; Wight, who
in his leones admits that he has not seen any of T ulasne’s
specimens, names the same plants Dalzellia or Tulasnea
pedunculosa and D. Lawii respectively. I have confirmed
this by actual examination of the specimens, which agree
with one another in detail. Lawia pedunculosa thus is a
mere synonym of L. longipes, and L, Lawii of L. pulchella.
212
WILLTS : PODOSTEMACEÆ
We now have L. foil osa, another of Wight’s species, to con-
sider. Specimens from his herbarium, named in his own
handwriting, are in the Kew herbarium; the packet is
marked on the outside “ Lawia No. 1, Law, Bombay,” in
Law’s handwriting. The character on which Wight bases
the species is the supposed absence of any union of the
leaves at the base of the pedicel of the flower ; in the other
species the leaves are always united more or less into a
cupular structure, or rather, the axis bearing the leaves is
cupular. Examining Wight’s material, I found the foliose
character in a few of the flowers, but others had a distinct
cupule, and in all other respects the material exactly
resembled his L. pedunculosa, which is also Law’s “ Lawia
No. 1 .” Wight, therefore, appears to have divided up Law’s
material into these two species on the chance observation of
one or two flower axes (probably supposing each axis to
represent a whole plant), and the fact that some axes were
foliose, others not, showed the species to be untenable on his
material. Warming, however, describes in detail a specimen
collected at Khandalla by Goebel, which has the foliose
character very well marked, at any rate in the bud stage,
which is all that Warming {1. c. IV., 159) figures. By the
kindness of Prof. Goebel, I have had the opportunity of
examining his material, and find that it is usually cupular,
but that in the bud stage it is occasionally foliose, and I have
further verified this observation upon living material
collected by myself at Khandalla and elsewhere in the Ghats,
east of Bombay. Wight’s L. foliosa, therefore, falls to the
ground as an independent species, becoming a syjmnym of
his pedunculosa, ?’.e., of Tulasne’s longipes.
We are thus reduced to Tulasne’s three species, which, so
far as herbarium material shows, are well separated, mainly
by length of leaves and fruiting pedicels. I have, however,
found by examination of large quantities of material that
these characters are too variable to be relied upon, even in
living material. Long and short leaves and pedicels may
OF INDIA AND CEYLON.
213
occur on the same plant, and it is by no means difßcult to
get from one plant in the Bombay Ghats pieces resembling
the types of L, pulchella and L. longipes respectively.
These two must therefore be united, and the only question
left is whether the Indian forms are specifically distinct from
the Ceylon forms. Enormous labour will be needed to fully
settle this question, but for the present I have no hesitation
in uniting them, though, as already mentioned, every locality
seems to have its own peculiar form. The Indian forms
agree fairly well among themselves in the following points
of difference from those of Ceylon : they have a rather longer
fruiting pedicel, a more obovoid and lighter coloured fruit,
thinner fruit ribs, and usually rather longer leaves. The
thallus type of the northern forms is more like that of the
common Ceylon forms than that of the southern, so that it
is possible that the two should be separated on geographical
grounds, but for the present our knowledge is insufficient.
Four groups of forms are found among the material at my
disposal.
S. W. India and Ceylon. Only species L. zeylanica.
LA WIA ZEYLANICA, Tul. in Ann. Sc. Nat., ser. 3, t.
XI., p. 112 ; Trimen, FI. Ceyl., III., 416 ; Warming, &c. ; L.
pulchella, Tul., 1. c. ; L. longipes, Tul., 1. c.; L. Lawii ( Wight),
W arming , O . Kuntze ; L. pedunculosa ( Warming,
O. Kuntze; L. foliosa {Wight), Warming, O. Kuntze;
Terniola zeylanica, Tul., Podo. Monogr., p. 190, t. 13 ; T.
pulchella, Tul., 1. c. ; T. longipes, Tul., 1. c. ; Terniolæ spp.
omn., ramosissima excepta, Wedd. in DC. Prodr., XVII., 46,
Hooker in FI. Br. Ind., V., 62 ; Tristicha ceylanica, GardnAn
Calc. Journ. Nat. Hist., VII., 177, 1846 ; Dalzelliæ spp. omn.,
ramosissima excepta, Wight, Ic. PI. As., p. 34 ; Thwaites,
Enum. PI. Zeyl., 223, C. P., 3,809 ; Tulasneæ spp. omn.,
ramosissima excepta, Wight, 1. c.,tt, 1,919, 1,920 ; Mnianthus
zeylanicus, pulchellus, longipes, Walp., Ann. III., 443.
Thallus frondose, creeping, with ribbon-like or flabellif orm
tips ; flowers solitary, very numerous, terminal, emerging
(^9)
214
WILLIS ; PODOS'L'KMACBÆ
from leafy axial cupules of varying depth ; leaves and
pedicels varying in length, apical leaves dimorphic.
Thallus frondiform, rootless, closely attached to the rock by root
hairs to the tips of the growing apices, often reaching 20-35 cm. in
diameter, spreading in all directions, about 0’5 mm. thick, hard, and
siliceous, deep crimson or green in colour when alive, much branched,
with flabelliform or ribbon-like apical lobes ; the branches to 10 mm.
broad, diverging at a moderate angle from the main stem, closely
crowded or scattered. Leaves very numerous, both at the growing
points, where they are borne directly on the thallus, and on the older
parts of the thallus, where they occur in endogenous rosettes, linear
to linear-ovate, acute, very thin and delicate, 1-15 mm. long, with
white mesial line in lower part, attached by broad base. Apical leaves
dimorphic, the lateral ones distichous on the edges of the thallus,
broader, the upper linear, more numerous, irregularly arranged on the
upper surface. Flowers very numerous, closely crowded along the
edges of the thallus, solitary, emerging from small terminal cupules,
rarely on upper surface. Cupules ovate, about 0 5-2-0 mm. deep
and 1‘0-L5 mm. broad at the mouth, axial, densely leafy above and
on the edges, with thick upper and thin lower wall, free from
thallus below, more or less ascending, bristly when dry. Peduncle
from base of cupule about 2-6 mm. long when the flower opens,
coated with pellucid cellular tissue which falls after flowering.
Flower erect on the peduncle, often only emerging from the cupule
when the water falls so as to expose the tip of the latter, which often
remains closed so long as submerged. Perianth segments united for
2/3 of their height. Ovary ellipsoid subsessile, stigmas linear.
Capsule about 2 mm. long, ellipsoid.
On smooth rocks in rapids, Ceylon, S.-W. India from
Travancore to Bombay, to 2,500 feet, common.
This is an extremely variable plant, and specimens from
many localities have been examined. Each locality seems
to have its own form, but after close examination of all
I incline to group them into four varieties, partly geo-
graphical.
L, zeylanica Qardneriana ; Tristichaceylanica, —
Thallus large, with flabelliform apices; leaves 1-3 mm.
long ; rosettes numerous. Cupules ovate, about l*5-2*0
mm. deep, shortly bristly in dry specimens. Pedicel
ultimately about 6 mm., long. Fruit ellipsoid, dark
brown.
OF INDIA AND CEYLON.
215
Ceylon only. First found by Gardner. Maliaweli-ganga
at Hakinda near Peradeniya, Gardner Î Thwaites (C. P.
3,809) ! Willis ! and at Haragama below Kandy, Thwaites !
Laggala-oya (tributary of Mahaweli-ganga) in Matale
east, Thwaites ! Guru-oya at confluence with Hulu-
ganga (tributary of Mahaweli-ganga) near Teldeniya,
Willis ! Kelani-ganga near Kitulgala, H. F. Mac-
millan !
L. zeylanica Parkiniana —
Thallus smaller than in last, more definitely branched,
with long ribbon-like apical lobes, each terminated by
one or few growing points ; leaves usually 3-6 mm.
long. Cupule with longer bristles. Fruit as in last.
Ceylon only. First found by Mr. J. Parkin at Hakinda
in the Mahaweli-ganga ! Guru-oya at confluence with
Hulu-ganga, Willis ! The latter specimens are rather
smaller in all respects.
L. zeylanica malaharica ; L. spp. indicæ, Auctt. —
Thallus to 6 inches or more in diameter, with frequent
branching and ribbon-like lobes, tending to flabelliform
lobing in some of the more northern forms. Leaves
3-9 mm. or more. Cupule usually short. Fruiting
pedicel variable, averaging about 8 mm. long. Fruit
obovoid-elliptical, 1-2 mm. long, light brown, with thin
ribs and marked depressions between the carpels.
Seeds light brown.
South India. Tinnevelli Ghats in the Tambraparni river,
C. A. Barber, S. India Flora, No. 2,847 ! South Kanara,
at Sullia, do. Nos. 2,150, 2,153, and at Beltangadi,"*^ do.
2,516, 2,524, 2,525 ! Anamalais (.^) Wight, spec, in hbb.
N. Kanara, Talbot in the Kala Nuddi, No. 1,129, mixed
with other Podostemaceæ. Wight and Talbot’s speci-
mens are too fragmentary for determination.
In streams above the bridge near the Dâk "bungalow. — C. A. B.
216
WILLIS : PODOSTBMACEÆ
L. zeylanica konkanica ; L. spp. indicæ, Auctt. —
Thallus to 6 or 10 inches, with fiahelliform lobes and
crowded growing points, occasionally tending to ribbon
forms with single or few growing points. Leaves
6-15 mm. Cupule short or long, often foliose when
young. Flowers often nodding in bud. Fruit pedicel
very variable, averaging 8-9 mm. Fruit 1-|-2| mm.
long, obovoid, light brown, otherwise as in preceding.
Western India. Igatpuri in the Thul Ghat,* Willis, 1,800
feet ! Kasara, Thul Ghat,t 800 feet, Willis ! Khandalla
and Lanauli,J in the Bhor Ghat, 1,500-2,500 feet,
common, Willis ! Sakarpathar, near Lanauli, Woodrow !
Tiger Leap, near Khandalla, Woodrow ! Khadshi river,
near Bhorkus, west of Poona, R. K. Bhide ! The
Konkan, Dalzell, Law, Ritchie, &c., in herbaria ; speci-
mens too fragmentary for determination.
EUPODOSTEIVIE/E.
DICRÆA (male Dicræia) (Du Pet. Th., Nov. Gen. Madag.,
p. 2, No. 4, 1806), Tul. in Ann. Sc. Nat., 3 ser., t. XI., 1849,
p. 100, and Podo. Monogr., p. 114 ; Wedd. in DC. Prodr.,
XVII,, p. 67 ; Wight, Warming, 11. cc. ; Podostemonis spp.^
Griffith in As. Res., XIX., 103 ; Gardner in Calc. J.N.H., VII.,
p. 179, and in Flora, 8, p. 40; Bentham and Hooker, Hooker in
FI. Brit. India; Trimen in FI. Ceyl., &c.; Poly pleur urn ^ Tayl.
MS. ex Tul., Warming, p.p. ; Lads (Lindl.), Steud., p.p.
Flower zygomorphic, naked. Stamens (2 or 1) with stami-
nodes on either side of common axis. Ovary smooth, ripen-
ing to 8- (rarely 8-12-) ribbed isolobous fruit with oo seeds.
* In a stream below the Bombay road on the right-hand side going from
Igatpuri, at about the 31st milestone.
f In streams crossing the road to Bombay below Kasara station, and
especially in the large stream about one mile down the road, at a series
of rapids about ^ mile down stream from the bridge.
% Enormous quantities in the stream below the Hamilton Hotel on the
left going from Khandalla to Lanauli ; also in stream on the right after
crossing the railway from the Hamilton Hotel, and in nearly all streams
in the district and on far side of Sakarpathar.
OF INDIA AND CBYDON.
217
Thallus various, usually freely floating from attached base
exogenously branched, with marginal ultimately 1-flowered
secondary shoots. Flowers enclosed in spathes, splitting
irregularly at the tip, and subtended by 2-8 (usually 4)
fleshy scaly bracts.
Flower zygomorphic, naked, enclosed before anthesis in a tabular,
usually oblanceolate spathe, which opens irregularly at the tip, pedicel-
late, the pedicel lengthening as the fruit ripens and shedding its
deciduous cortex. Stamens 2, rarely 1, monadelphous, with a
filamentous staminode on either side of the common stalk. Pollen
didymous. Ovary symmetrical, elliptical, 2-locular, with two equal or
unequal subulate marcescent stigmas with small papillae. Ovules oo.
Capsule isolobous, 8- or rarely 9-12-ribbed, the ribs on both valves
decurrent into the pedicel, septifragal, both valves persistent. Seeds oo .
Submerged herbs with the habit of Fucus and other seaweeds.
Primary axis (? always) very short, non-flowering, giving rise laterally
by endogenous development to a thallus of phylogenetic root nature,
exogenously branched with root cap, ribbon-like, cup-like, filamentous,
f ucoid, often crisped or twisted, attached to the rock by a foot or by
haptera, or by a creeping basal portion, or at all or most points, but
usually with the distal parts drifting freely out in the water. Secondary
axes CO , endogenous on the upper sides of the thallus near the edge, or
rarely in the central parts, consisting in the vegetative season each of a
fascicle of small leaves with included evanescent axis, and all or some
of them ultimately floriferous. Vascular bundles leading to floriferous
shoots, and immediately adjacent parts of tissue of thallus becoming
woody in flowering season, the rest of the tissue and the non-flori-
ferous parts ultimately falling away (as, e.g., in most herbarium
specimens). Floriferous axes exserted, with 2-8 (usually about 4)
distichous imbricated bracts, the upper larger, narrowly linear to broadly
ovate or helmet-shaped, sheathing, thicker on the upper side, formed
by the enlargement of the sheathing bases of the leaves and the fall or
decay of the tips. Flower solitary, terminal, enclosed in spathe, opening
when exposed to air.
The Indian and Ceylon species here included form a very
natural group. There are also two species in Madagascar,
about which more information is required. I use the genus
in practically the same sense as defined by Tulasne. The
latter’s section Ceratolacis was made into an independent
genus by Weddell, as mentioned above, but on further exami-
nation may probably be found to be best placed in Dicræa.
Tulasne’s other sections are based on insufficient knov/ledge
218
WILLIS : PODOSTEMACBÆ
of the forms, and the grouping of the Dicræas under Podos-
temon, as in Bentham and Hooker’s Genera Plantarum and
in the Flora of British India, is still more artificial, com-
pletely breaking up the genus and including in the section
Dicræa the markedly different form Podostemon subulatus.
Weddell spoiled the genus by admitting the smooth
anisolobous-fruited Willisia selaginoides.
The Dicræas are very abundant in all the Podostemaceæ
regions of India and Ceylon, and are extraordinarily variable
and polymorphic. It is not yet possible to draw good specific
limits, and unsafe to try to define varietal forms. I give
below a tentative grouping of the forms studied by myself
in fresh and preserved material, and have drawn the specific
limits very widely.
Owing to the curious way in which the parts of thethallus
not directly concerned in the production and nutrition of the
flowers and fruit break away towards the end of the life of
the plant, there is a great difference in the morphological
structure, as already mentioned (p. 191), between the sub-
merged plants and the dry fruiting specimens, in most cases.
This will be more clear after the appearance of the figures
illustrating the genus in the subsequent paper.
We have so little knowledge of the Madagascar species that
it may easily prove to be the case that they are generically
distinct, in which case our genus will have to be re-named.
Madagascar, Ceylon, Travancore, and Anamalai Mountains
to Kanara, Khasia Mountains of Assam, Burma. Abundant
in all districts mentioned, variable.
Dicrœœ indicœ.
Thallus narrow, long, with creeping basal portion and
free distal. Secondary shoots (floral buds) sessile on thallus
by narrow base and tapering downwards, or stalked, often
linear or narrow oblong.
Thallus cylindrical, filamentous, to 50
cm. long in free parts ; pedicel of
fruit about 8 mm. ...D. elongata, Tul.
OF INDIA AND CKYLON.
m
Thallus narrow, ribbon-like, flattened,
to 30 cm. in free parts; pedicel about
6 mm. ... ...D. dichotoma, Tul.
Thallus various, usually broad, rarely over 20 cm. long.
Secondary shoots (floral buds) sessile on thallus by broad
base, and often square, or broader than high ; bracts usually
markedly helmet-shaped.
Pedicel of fruit under 5 mm. long. ...D. minor, Wedd.
Pedicel of fruit over 5 mm. long.
Thallus usually small, to 10 cm.
long and 1 broad, algiform,
branched, but becoming appar-
ently much branched and narrow
in dry specimens by breaking
away of intermediate tissues.
Pedicel 8-25 mm. N.E. India and
Burma ... ...D- Wallichii, Tul.
Thallus usually large, algiform,
often to 30 cm., usually much
branched. Pedicel 6-40 mm.
S.W. India and Ceylon ...D. stylosa, Wight.
DIGEÆA ELONGATA, Tul. in Ann. Sc. Nat., 3 ser., t.
XI., 1849, p. 102, and in Pod. Monogr., p. 124 ; Wight,
Thwaites, Wedd., Warming, W. cc.; Podostemon elongatum,
Gardner, in Calc. J. N. H., VII., 1846, p. 188; Hooker, Trimen,
11. cc.
Thallus long, narrow, cylindrical, tapering to filamentous,
attached at base by creeping hapterous part, and with
cæspitose freely floating parts to 50 cm. long. Secondary
shoots distichous, the lower to ultimately floriferous,
the rest deciduous with tips of thallus. Floral buds usually
slightly stalked, narrow. Fruiting pedicel about 8 mm.
Thallus cylindrical, dimorphic, rarely over 3 mm. thick in any part;
lower part frequently branched, creeping, attached by haptera and
root hairs, and giving off cæspitose freely floating branches to 50-60
cm. long, sparingly branched and tapering to capillary ends. Secondary
m
WILLI8 : PODOSTEMACEÆ
shoots distichous, chiefly on the floating thalli, usually about 5 mm.
apart ; the lower or f or rarely more ultimately floriferous, Flori-
ferous part of thallus woody, and the rest of the thallus, including
the marginal parts of the floriferous portion, usually deciduous, at least
on exposure to the air. Floral shoots in bud usually linear to oblong,
widening upwards, not crowded together, the axis elongated above the
surface of the thallus, with 2-5 (usually 4) distichous scales. Scales
imbricate, often connate at base, most often with pointed tips after the
fall of the leafy ends, but frequently more or less cowl-shaped. Spathe
at first included, soon emergent, 2-5 mm. long when open. Pedicel
about 4-5 mm. lengthening to about 5-9 mm. in ripe fruit. Stamens
2, equalling ovary with stigmas, staminodes shorter. Ovary smooth.
Stigmas 2, sessile, shorter than ovary, stout, subulate, marcescent or
deciduous. Capsule 1-24 mm. oblong- elliptical, 8-ribbed. Fruiting
thallus more or less erect on rocks.
Ceylon on rocks in rapids to 2,000 feet. Flowers January
to March. First discovered by Gardner. Endemic.
Mahaweli-ganga in Ambegamuwa, Gardner, Mahaweli-
ganga near Peradeniya, Gardner ! Thwaites, C. P. 2,259 !
Trimen ! Mahaweli-ganga at Hakinda rapids, Willis !
Kelani-gangaatKitulgala, Trimen ! Bambarabotuwa-ganga,
H. F, Macmillan !
A very distinct species, easily recognized by the long fila-
mentous fioating thalli, and by the comparatively widely
separated and narrow fioral shoots. The thallus tips are
commonly missing in herbarium specimens. The Bambara-
botuwa specimens have a rather smaller habit and more
cowl-like bracts, but more material is required to decide
whether they are specifically or even varietally distinct.
Sometimes mixed in herbaria with D. stylosa laciniata.
Good figures of this species are given by Tulasne (Monogr.,
t. IX.) and Warming (1. c. II., t. X.). See also PI. XVIII.,
XIX.,. in subsequent paper.
D1GRÆA DICHOTOMA, Tul. in Ann. Sc. Nat., 1. c., 1849,
and in Mon. Pod., p. 119 ; Wight^ Wedd.^ Warming^ II. cc.;
D. Wightii, Tul.^ 11. cc.; D. rigida, Tul,, WiglityW.cc,:
D. longifolia, WightA-Q>\ Podostemon dichotomum, Wightii,
rigidum, Gardn, in Calc. J. N. H., YIL, 1846 ; P. dichotomum,
vars., Benth. in Hook. FI. Br. Ind., &c.
OP INDIA AND CEYLON.
221
Thallus long, narrow, thin, ribbon-like, oval in section,
often zigzag, attached at base by creeping portion or foot,
and with freely floating parts to 30 cm. long. Secondary
shoots distichous, the lower ultimately floriferous, the
rest deciduous with the tip of the thallus. Floral buds often
slightly stalked, narrow. Pedicel of fruit about 6 mm.
Thallus flattened, oval in section, to 6 mm. wide, tapering ; lower
portion usually sparingly branched, creeping, attached to rock at base
and usually at several outer points, upper usually freely floating, to
10-30 cm. long, frequently branched, linear, often zigzag. Secondary
shoots distichous on all parts of thallus, about 2-3 mm. apart, often
on small lateral projections of the thallus, the lower or almost all
ultimately floriferous, and then elongated above the thallus much as
in the preceding species ; non-floriferous parts of thallus usually
deciduous on exposure. Floral shoots in bud not closely crowded
together, oblong or lanceolate to obovate, tapering downwards.
Scales 4-5 or more, lower often connate, more or less helmet-shaped,
at first with elongated leafy tips, which sometimes persist even after
exposure (Z). Wightii phase). Spathe much as in preceding species.
Pedicel 3-6 mm. in flower, elongating to about 6-7 mm. in fruit.
Stigmas ovate to subulate, marcescent or deciduous. Fruit 1-2 mm.
long, 8-ribbed, the dehiscence ribs broader than in D. elongata.
Fruiting thallus usually prostrate on rocks, with erect fruit stalks.
Nilgiris and Malabar Hills, on rocks in rapids, to 6,000
feet. Flowers November to March. First discovered by
Gardner and Wight.
Paikara river, Nilgiris, near Paikara, Gardner ! Wight Î
Proudlock ! Barber ! Gamble, 11,746 ! Willis River
near Bangi Tappal,f Nilgiris, Gamble, 13,333 ! Malabar,
Rev. E. Johnson !
A distinct species, fairly easily recognized by the thin
ribbon-like thalli, short pedicels, and narrow floral shoots.
It most nearly resembles D. stylosa, var. laciniata, or D.
elongata.
Gardner, who originally discovered and described this
species, divides it into three, to which Wight adds a fourth,
* Abundant at all ra^pids, commencing below the bridge near the Dak
bungalow at Paikara ; especially good at Paikara Falls.
t About 22 miles south-west of Ootacamund. near Avalanche.
(30;
222 WILLIS : PODOSTBMACEÆ
D. longifolîa. Tulasne accepts these with hesitation, and
Weddell, followed by Hooker, places them all under D.
dichotoma as varieties. Without much further study, I am
not prepared to admit even the varietal rank of these forms.
The thallus varies much in stoutness and width, and in the
degree of zigzagness, characters on which a good deal is
based in the separation of the species mentioned. The
form Wightii is mainly characterized by the apparently
greater number and length of the bracts, but this depends
merely on the specimens having been taken before the
fall of the tips. The form longifolia is separated from
dichotoma mainly on length of the leaves, a very variable
character in most of the order. The species is figured by
Tulasne, Monogr., t. IX.
DIGBÆA MINOR, Wedd.in DC. Prodr., XVII., 1873, p
71 ; Podostemon minor, Benth, and Hk. /. ; Hooker in FI. Br.
Ind.
Thallus insufficiently known, apparently creeping,
branched, ribbon-like, to 10 mm. wide and to 5 cm. long.
Secondary shoots when floriferous sessile by broad base on
thallus margins ; bracts helmet-shaped. Fruiting pedicel
very short, usually under 3 mm.
Thallus (c/*. Griffith’s rough sketch in Hb. Kew) stellately branched,
algiform, ribbon-like, creeping, to 5 cm, long from centre and 3-10
mm. wide, hapterous at various points on the lower side. Secondary
shoots marginal as in other species, some (all ?) ultimately floriferous.
Bracts 2-5, more or less helmet-shaped, keeled, like those of D.
Wallichii, or occasionally (?) like those of D. elongata, imbricate.
Spathe very short, funnel-shaped. Pedicel of flower about 1 mm.,
lengthening to about H-4 mm. in fruit. Capsule 8-ribbed.
Assam, in the Khasia Mountains,
Borpani river, north of Nartiang, near Madan, Griffith,
No. 2,437, ex. Hb. E. Ind. Co.; Khri river, below Nongkhlaw,
Hooker (spirit material in Kew Museum, det. Weddell, fide
Oliver).
This species is very imperfectly known, and may be only
a phase of D. Wallichii. Griffith’s specimens in the Kew
OF INDIA AND CKYLON.
228
Herbarium are mere fragments. The spirit material collect-
ed by Hooker may be a different species or form ; I am not
able to decide this question on account of insufficient supply
of material. The main, or almost only, distinguishing mark
of the species is the short pedicel of the fruit, but this may
be due to its having been gathered unripe {cf. p. 194) or other
cause. Fresh investigation is required on the spot where
Griffith found it ; I much regret that I was unable to visit
the place. The capsule is 8-ribbed,but Weddell describes it
as having 10 ribs, probably having examined one in the early
stages of opening. The idea of a sheath {cf. Weddell) at
base of the flower stalk arises from the examination of bad
herbarium specimens. The pedicel is so short that the bracts
separate mainly on the top side to allow the escape of the
flower.
DICRÆA WALLICHIf Tul. in Ann. Sc. Nat., 1. c., p,
101, and Podo. Monogr., p. 118; Wight, Wedd., 11. cc.;
Podostemon Wallichii, R. Br, ex Wall., Cat. No. 5,225 (1828),
Griff, in As. Res., XIX., 103 (descr. et ic.), Gardn., Benth.,
B.ooker,\\.ca. ; Podostema Wallichii, Royle, 111. Bot. Himal.,
I., 331 ; Dicræa pterophylla, Wedd. in DC. Prodr., XVII.,
71; Podostemon pterophyllus, Bentli., Hooker in FL Br.
Ind.; Lacis Wallichii, Polypleurum orientale, Tayl.
MS. ex Tul.; P. Wallichii, Warming, 1. c. VI. ; Blandowia
striata, Lehm. MS. ex Tul. (non Willd.).
Thallus algiform, ribbon-like, to 10 mm. wide and 10 cm.
long, branched, more or less creeping, and attached at base
and other points. Secondary shoots marginal or rarely
central, some or all ultimately floriferous. Vascular bundles
leading to flowers, and immediately surrounding tissue
woody, the rest with the non-floriferous part of thallus
deciduous. Floral buds sessile by a broad base, bracts
helmet-shaped. Pedicel of fruit 8-25 mm. N.E. India and
Burma.
Thallus Fucus-like when alive and submerged, attached at base by
a foot and on outer parts by haptera, prostrate, branched, and lobed.
224
W T 1 .LIS : PODOST KM A (J KÆ
to 10 ciTi. long or more and 1 cm. wide, ribbon-like. crisp, fleshy, with
broad blunt apices ; when gathered dry and in ripe fruit, small with
cup-like base and fimbriate fruit-bearing outer ends. Secondary
shoots on upper edges of margins, usually 2-3 mm. apart and hence
rather crowded in flower, rarely on central parts of thallus, all or some
ultimately floriferous, and then closely sessile on thallus and as broad
as long before emergence of spathe. Bracts 2-7, usually 4, markedly
cowl- or helmet-shaped, keeled, usually divaricate when open.
Vascular bundles leading to flowers, with immediately adjacent tissue
becoming woody, the rest of the thallus tissue deciduous, thus much
altering the habit of the fruiting specimens. Spathe oblanceolate to
6 mm. long, splitting irregularly at the tip or occasionally on the under
side. Pedicel of flower about 4-5 mm., elongating in fruit to 8-25 mm.
Stamens equalling ovary and stigmas. Stigmas subulate, unequal, the
lower larger. Capsule 8-ribbed, or with 2-4 extra ribs intercalated ;
ribs very narrow.
Assam and Burma on rocks in streams to 4,500 feet.
Flowers October to February.
This species probably may be divided into two or more
varieties or species. Wallich’s original plant, found by
Gromez in the Sylhet Hills, and all the Cherrapunji speci-
mens of D. Wallichii apd D. pterophylla, so far as I am
aware, have short pedicels and eight ribs to the capsule. Some
of Griffith’s Assam specimens, locality unknown, have longer
pedicels and more ribs than eight, and come near the speci-^
mens collected by Lehmann at Chepedong near Moulmein in
Burma, and those collected in the same district by Parish
in 1859. As mentioned above, D. pterophylla cannot be
regarded as a separate species, being founded on material
differing only from the typical D. Wallichii in having the
keel of the bracts less disintegrated. I incline to think that
there are two good varieties, geographically separated by the
great valley of the Surma.
D. Wallichii Khasiana ; D. Wallichii (R, Br.), TuL;
Wedd. ; D. pterophylla, Wedd,
Pedicel of fruit 8-15 mm. Capsule 8-ribbed.
Khasia Mountains, plateau of Cherrapunji, 3,000-4,000 feet,
on rocks in streams running towards Maosmai and
OF INDIA AND CEYLON.
225
Maoraloo, Griffith! Hooker! Clarke! Willis!'^ Sylhet
rivers, Wallich ! Sylhet Mountains, Gomez !
D, Wallichii striata; D. Wallichii {R. Br.), TuL, p.p.;
Wedd. p.p.; Blandowia striata, Lehm. MS. ex Tut.
Pedicel usually 10-25 mm. Capsule usually with more than
eight ribs.
Mountains of Burma (and Assam ?), Chapedong near
Moulmein, Lehmann, No. 3, 1- CO
p 'îK P
Ö CO CO
O
1
Castilloa
creamed.
273 grm.
100 per
cent.
91*78
7*54
0*50
0*18
100*00
6*38
6*34
8
Para,
Mercuric
Chloride,
1899.
266 grm.
98 per-
cent.
94*39
5*30
0*21
0*10
100*00
3*17
3*00
9
Para,
Cold
Acid
and
Creosote,
1899.
452 grm.
98 per-
cent.
94*14
5*62
0*14
0*10
100-00
2*88
2*89
10
P ara.
Hot
Acid,
no
Creosote,
1898.
4:47 grm.
100 per
cent.
93*43
5*86
0*49
0*22
O O .-H
p p
Ö ÔQ ca
o
3
Para,
Hot
Acid
and
Creosote,
1898.
224 grm.
100 per
cent.
94*11
5-47
0*21
0*21
100 00
2-92
3*06
11
Para
Ball,
1897-98.
22
Years.
g . O fN O CO
X P 43 O O CO CO
he —I a gn iP .L, Ô
zo O
CO
CO
9Z-f
9Z-f
00-001
5
Para
Ball, !
1897-98.1
12 1
Years. ‘
g p C3 «0 05
fH 2 42* CO 05 P vH
&r '-'a lb Ô
o o o
O tr^
Ô M PÏ
o
!
4
Para
1 Shell,
1898.
j 22
' Years.
1
g P4 . 05 p. p
2 C2 ,-1 p CO
^ J g à Ô Ô
^ p 2 05
r5
42 a
CO O
^ Ü
323
0) 42
I'S
si
ap
P 03
>03 a
^ P
o S oE
^ > a .
o3P’n O
Ä ° g o
3§'s|
42 ÖJ0 ^ cc
O O 05 42
bfl f.. ^ 05
2-p 2
!=> a'S
S ^ 2 a
Orö ^ So
. . .
o ;>>-^COCOlC
lllllgl
o ^ O w zn a:
m
Il 2
tS'S ^
2-So.^ g
i rl 12
a CO p
2 9 9 p
iS 2 ^ a
■ë^'3-S 8
ci 2 ® 8
■■2 05 05
&
^ sT λ 4^ P
ob^o .2
*
i ^ “p S
s,l|g s
T 0
a i
p^pqpn <3o3
25e;
^^OTBS.
Preparation of the Sample. — The total weight was taken, and about
half of each sample, after exact weighing, was reduced to a thin
uniform sheet under a stream of water in the usual machine employed
for the purpose in the rubber industry. The film thus obtained was
dried in a dark chamber at 30-35° C. and weighed, (Specimens may
be seen in the Peradeniya Museum.) These films were then used for
the further analyses.
Estimation of Moisture. grammes of the film were weighed
between tared watch-glasses and dried for 6 hours at 60° C., after
which the weight is constant, but when exposed to ordinary air increases
by the 0’3-0'5 per cent, of moisture, which is normally present in
caoutchouc.
Estimation of Ash. — An exact weight, about 2 grm., was calcined
in a platinum crucible, the operation being carried on very slowly until
complete distillation of the caoutchouc, when a little ignition yields a
perfectly white ash.
Estimatioîi of Resins soluble in Acetone. — An exact weight of about
2 grm. of film was extracted in a Soxhlet’s apparatus. The form used
was a globe of about 100 cc., containing about one-half its volume of
pure anhydrous acetone, CH3 (COOH) , The film after extraction by
four hours’ boiling was dried in vacuo. The loss of weight, less the
moisture-content determined by the first analysis, gives the proportion
of resin. The figures obtained in different analyses were very con-
sistent, as the table shows.
Estimation of Resins soluble in Mixture of Benzine and Alcohol. — Five
grm. of film were dissolved in 100 cc. of pure benzine and 150 cc. of
strong alcohol (95 per cent, or stronger), then stirred in. The caout-
chouc is precipitated as flakes, which run together into a clot ; this is
washed several times with alcohol, and dried till the weight is constant
on a glass or porcelain dish in a chamber kept at 30-35° C. The loss
of weight is regarded as representing the total resin. The repetition
of this operation gives a further loss of 0*5 to PO per cent, of the
weight of the caoutchouc, and a third operation a further one of
usually about 0'25 per cent. The figures given in the table are the
result from one operation only and therefore require correction, but
are comparable among themselves.
We have now to consider what conclusions may be drawn from these
analyses.
In the first place, it is evident that the cleanliness of the rubber
prepared by Mr. Parkin’s methods is very great. Most commercial
Para rubbers after treatment with the machine yield only 84-90 per
cent, of rubber, but Mr. Parkin’s give at least 98 per cent., and some
100 per cent. Hence the rubber thus prepared should obtain, other
things being equal, at least 10 per cent, higher price. In actual fact
some of the rubber lately exported from Ceylon obtained 4s. 2d. per
lb. against 3s, 96?. for best native Para, figures corresponding very
NOTES^.
257
closely to the above estimate. MM. Michelin state that the quality
of the samples sent to them was at least equal to that of the very best
corresponding commercial rubbers, but that an exact valuation could
not be given, as the samples were too small.
A comparison of Mr. Parkin’s samples with the samples obtained
by the old method of tapping shows that the shell rubber of the latter
is almost or quite as good as regards the yield of rubber from the
sample. It is thus evident that the great advantage in all these
processes is their cleanliness in collection of the milk ; comparison
with the ball rubber of the old process shows the difference. There is
no special virtue, as regards the percentage of rubber obtained, in one
method rather than the other. In actual fact there is no great gain
in using the acid if the climate be fairly dry, so that the natural
coagulation and drying can take place quickly ; but the use of creosote
as a preservative is perhaps always advisable.
The acid or mercury coagulation method, however, seems to have
some advantage when we go on to the third set of figures giving the
detail of analysis. The greater the proportion of resin in rubber the
less its value, and the acid-prepared samples seem to have less than
the samples prepared by simple drying. MM. Michelin remark that
the proportion of resin soluble in acetone in the Ceylon samples is
closely similar to that in the corresponding native rubbers. They
remark also that the proportion of resin soluble in acetone or in benzine
and alcohol is no greater in rubbers altered by heat or light than in
those not so altered ; the proportion is only increased when there is a
chemical action or oxidation at the same time. Thus, sample 5, though
sticky, contains less resin than samples 3, 8, and 9, which are far
superior for commercial purposes. Caoutchoucs may be quite spoiled
by the action of heat without increasing the proportion of resin.
Another point of interest is the difference between the rubber of old
trees and younger ones, e.g.^ as indicated in samples 2 and 4, in which
trees of 12 and 22 years old respectively were employed. It would
of course be absurd to lay much stress on these figures, but so far as
they go they show that the rubber of the old trees is no better, if,
indeed, it be not worse (as containing more resin) than that of the trees
12 years old.
From the planter’s point of view the chief lesson to be learnt from
Mr. Parkin’s work and the above analyses is to practise the greatest
possible cleanliness in collecting and preparing rubber ; any rubber
dried on the tree or otherwise contaminated will sell for a much lower
price as “ scrap ” or “ negro-head.” By Mr. Parkin’s method of
collection in separate tins under each cut and placing water in the tins
the great bulk of the milk can be brought in in the liquid condition and
filtered, and can then be treated in clean tins, with or without acid, as
local experience shows to be suitable. Quick drying is advantageous,
but excessive heat must be avoided, as under its influence the rubbei'
258
NOTES.
becomes sticky and loses most of its value ; sample 5, which was dried
on the tree in more or less sunlight, was thus deteriorated.
In conclusion, the opportunity may be again taken of calling
attention to some misprints which unfortunately escaped correction in
Mr. Parkin’s paper quoted (Circ. E. B. G., I.). On page 148, bottom
paragraph, the figures should read 0*9, 3’9, 0'25, and 8‘0 respectively,
and on page 152, paragraph 2, 0-83 per cent, mercuric chloride instead
of 8-3 per cent.
J. C. WILLIS.
Not® on the Product of the Latex of the Jak
(Artocarpus integrifolia).
As is well known, the jak tree exudes when tapped a large quantity
of a sticky milk, which hardens into a brittle substance. Inquiries are
frequently made as to whether this substance has any value for rubber
purposes. It is so brittle that it is evidently of no use for any of the
ordinary purposes for which rubber is employed, but it seemed to me
that it might have a value for ebonitising or such purposes. With this
in view I sent samples to MM. Michelin et Cie. of Clermont Ferrand,
who have kindly reported on it as follows : - “ A whitish resin, con-
taining 1‘8 per cent, of moisture. Melts completely in boiling water,
dissolves in benzine without previous swelling ; on addition of alcohol
the products precipitated in small quantity from this liquid do not
agglomerate. On a platinum foil it disengages an odour of burnt
bread. These characters indicate the absence of caoutchouc or anal-
ogous bodies (gutta-percha or balata). We do not think that this
product is of any value in the rubber industry.”
J. C. WILLIS.
Pith for Microscopists.
In Ceylon microscopists have hitherto relied, almost solely, upon
the pith obtained from the Elder and imported into this Island. It has
been a common experience when preparing serial hand sections to find
the local supply of pith exhausted, and this has invariably resulted in
an unsatisfactory series being prepared. The commercial material used
for making topees is too soft and fibrous, and one has had to fall back
upon the tubers of Solanum and Helianthus in cases of emergency.
When I was up-country a short time ago I came across cultivated
specimens of Fatsia papyrifera, Hook., known also as Aralia papyrifera,
or the Japanese rice paper plant. This plant yields an excellent pith,
which has long been used in China and Malaya for making paper and
artificial flowers. Though introduced into Ceylon about 1856, its
NOTES.
259
usefulness to microscopists seems to have been overlooked. The plant
will thrive at all elevations between 1,000 and 5,000 feet, and can be
easily propagated from cuttings. When once established it spreads
very rapidly, and a few well-established plants will yield a supply of
pith sufficient to meet the requirements of a large laboratory.
The pith is firm, white, and free from vascular bundles. The paren-
chyma cells are relatively large and the walls are very thin.
It can be obtained in quantity from the old and young stems, and
very often attains a circumference of 130 mm. In the young stem
the pith is solid to the centre, but in older parts the central area is
composed of thin parallel diaphragms of parenchyma alternating with
small air chambers.
The petioles and roots do not yield large quantities of pith, the former
being hollow and the latter nearly solid (vascular) to the centre.
I am informed that it is also particularly suited for entomological
Avork. Its firmness places it before Helianthus tuberosa, and its homo-
genous texture and pure white colour render it a much more valuable
pith than that from Elder for mounting very small insects,
HEKBERT WRIGHT.
Notes of Indian Travel, b^f a Ceylon Botanist.
Descriptions of the Indian flora and vegetation have usually been
given from the point of view of a botanist fresh from the temperate
zone. It is much to be desired that more work should be done in
India by visiting botanists — there is nowhere a larger or more varied
field for study, travel, and investigation. It is true that in certain
seasons the heat makes scientific work difficult, but there are many
cool hill stations, and the floras of the hill regions are of special
interest. It may be of use to future visitors and workers in the Ceylon
Botanic Gardens to give a few notes of travel in India in the cold
seasons of 1900-01 and 1901-02, in the course of which I visited a
considerable number of mountainous districts in my search after the
Indian Podostemaceæ, and to give these impressions in terms, not of
the very different European flora, but of the comparatively similar
flora and vegetation of Ceylon, which is known to many botanists.
Ceylon, by reason of its geographical position and good climate, with
the conveniences of travel and study which it possesses, forms a good
centre for preliminary work to be followed by excursions into India.
Travel in India is now a comparatively simple matter. In some
districts the resthouses (dâk bungalows) are furnished, in others not ;
in the latter case it is necessary to take furniture, and in all cases it is
well to carry supplies, as but little can be purchased in most villages.
260
NOTi^S.
A good travelling servant is of course a necessity in nearly all districts,
as no English is spoken even by the bungalow servants. Hotels are
to be found in most of the larger towns, and thé inclusive charge is
usually Rs. 5 per diem. Travelling in the country districts is very
cheap, especially in the south. Most parts of India are now accessible
by rail in a short time, and the fares are extremely low.
My first tour was from Bombay to Colombo, by way of the Ghats,
Poona, Bangalore, Madras, Ootacamund, and the Anamalais. My
second was from Calcutta into the Khasia Mountains of Assam and
into the Sikkim Himalaya. The two thus cover some of the most
botanically interesting districts of India.
The flat low-country of the Konkan and the lower western slopes
of the Ghats east of Bombay resemble the country between Colombo
and Kandy, but the vegetation is far less rich, though of the same
type. Many familiar Ceylon trees still occur, but the rainfall dis-
tribution is very difterent, an enormous precipitation in the summer
monsoon and a drought from November to May. The more Malayan
types of vegetation seem to disappear gradually, as the climate in going
northwards from Galle to Bombay changes from the well-distributed
rainfall and constant humidity of the former to the violent contrasts
that mark the latter. Distribution of rainfall appears to be of much
greater effect than total rainfall ; the total in the Bombay Ghats is
very high indeed. The difference also shows in the cultivated palms
on the coast, which in the Bombay District are mainly palmirahs, not
cocoanuts. The general effect of the whole district in December is
one of great dryness, only to be matched in Ceylon in a few districts
of the north and south-east. All exposed places have a brown or
yellow colour from the dry grass.
The configuration and scenery of the western side of the Ghats is
not unlike that of the Adam’s Peak region, but far less rich, and with
more precipitous valleys ; many of these are almost canons, with the
streams pouring into their heads over lofty precipices. They are
themselves usually well wooded, but the tableland above, which is at
an average level of about 2,500 feet, is mostly open grass country,
intersected by rocky stream beds, which in December are almost all
dry. I stayed at two centres for the study of the rivers, Igatpuri at
the top of the Thul Ghat on the Calcutta Railway (good furnished
bungalow) and Khandala at the top of the Bhor Ghat on the Madras
line (hotels). The former is very dry, and the lower slopes of the
Ghat are covered with teak forest. The latter is an excellent head-
quarters for the study of the flora, and though the open tableland is
no richer than the dry Uva patanas, which it much resembles, the
valleys contain an interesting jungle flora. On the eastern side of the
watershed the land slopes away gradually into the tableland of the
Deccan, with quiet rivers meandering through irrigated cultivated
country of little botanical interest. The chief tree is the Babul,
NOTES.
261
Acacia arabica. The wild flora is very xerophytic, the district receiving
very little rain. The heavy fall in the summer monsoon is only on the
western slope of the Ghats, the eastern side remaining dry, like the
Uva patanas in the south-west monsoon. At Poona there are some
interesting gardens, and there is a herbarium of the Bombay flora at
the College of Science under charge of Prof. Gammie.
The dry country continues all the way to Bangalore by the S. M.
Railway, but in one or two places, e.g.^ near Londa, the line passes into
the hilly district of the Ghats, and there is a somewhat richer flora
with stretches of jungle. At Watara station on this line the road to
Mahabaleshwar, the Bombay sanitarium, branches off. This place lies
at about 4,500 feet at the top of the Ghats, has an enormous summer
rainfall, and is said to be an excellent centre for the study of the
Bombay flora. I did not myself visit it.
From Bangalore, which is an interesting place and has a good
Botanic Garden, there is a considerable descent into the plains of
Madras, the general effect of which is similar to that of the cultivated
Tamil country of Jaffna in the north of Ceylon; but there are several
ranges of hills, which are covered with a scrubby jungle not unlike in
general appearance to that on the hills between Nalanda and Dambulla.
Madras itself is of no special botanical interest. The headquarters
of the Government Botanist are at Ootacamund in the Mlgiri Mountains,
to which I next proceeded. The ascent is very interesting, by a steep
cog-wheel railway, leading from the hot steamy valley of Metupalaiyam
to the little sanitarium of Coonoor at 5,500 feet. The flora at the
foot of the Ghat has the general aspect of that of the North Matale
District, and gradually changes on the ascent to a flora not unlike that
of Nuwara Eliya ; there is much less cultivation on these slopes than
in Ceylon. From Coonoor to Ootacamund the gradual ascent of 2,000
feet is over open rolling hilly country, cultivated in the hollows by
the hill tribes. Ootacamund itself, being in the centre of a large
plateau, has a comparatively dry climate, the violence of the north-east
monsoon spending itself further east, and that of the south-west mon-
soon further west in the Kunda Hills, which drop suddenly to the low-
country at the Western edge of the plateau, just like the Ceylon
mountains at Horton Plains. In all directions round Ootacamund
the country consists of open patanas, like the Uva patanas as seen
from Hakgala, but at a higher level. The flora is that of the drier
parts of Horton Plains, the jungle being found only in sheltered
valleys. To the botanist familiar with the up-country flora of Ceylon,
that of the Nilgiris is of singular interest, the genera being mainly the
same, but the species commonly different. All these mountain top
floras of the south, including those of the Nilgiris, Anamalais, and Palnis,
require much further investigation and comparison. Ootacamund is a
particularly good station for such work, and there are good resthouses
at various places on the plateau ; the climate is perhaps the finest in
(35)
262
NOTES.
the East. There is a pretty little botanic garden in the town, and the
Government Botanist of Madras also has his headquarters here, with
herbarium and laboratory.
After leaving the Nilgiris I visited the not less interesting range of
the Anamalais, on the other side of the great Palghat gap. This
great mountain range is only now being opened up, and it is still
necessary to take all camping materials in order to visit it. Leaving
Podanur station, it is about 33 miles to the foot of the hills at Vanan-
thorai, whence a cart road goes up to the planting district, which
occupies a portion of the lower central plateau from 3,000 to 5,000 feet
in elevation. The ascent is through forests of the type seen about
Nalanda and other parts of the north-eastern sides of the Ceylon hills.
At the summit level one passes into an evidently wetter region, and
the central parts of the range have a very rich and interesting flora.
They resemble the Ceylon hills in contour and vegetation, and are
very little interfered with by any cultivation. Range after range of
rolling hills can be seen, covered with high forest, intersected by clear
streams, and broken by stretches of patana on the higher ridges, some
of which reach to a height of close on 9,000 feet. The trees in the wet
jungles up to 4,000 feet are remarkable for their great size, larger than
those of corresponding elevations in Ceylon. There is an ample field
for work in the study of the formations into which this flora groups
itself ; preliminary observations might well be made in Ceylon, where
travelling and other facilities are greater, but the Anamalai flora is
finer for actual work of this nature, being so much less interfered with
by man. By working eastwards across the range the sister group of
the Palnis, with a drier climate and greater patana area, may easily be
studied, and accommodation can be obtained at the little sanitarium of
Kodaikanal.
The journey from the foot of the hills to Tuticorin is through well-
cultivated country of the general type of the Jaffna District of Ceylon.
My second tour was commenced from Calcutta, an interesting centre
for the botanist from its magnificent botanic gardens and economic
museum. It possesses, however, no laboratory accommodation. I first
visited the Khasia Hills of Assam, described by Sir Joseph Hooker,
much of whose track I followed. The hills can be reached in two ways
by steamer from Goalundo, either up the Brahmaputra Yalley to
Gauhati or up the Surma Valley to Companygunj. The boats are
comfortable and the journey easy. I myself followed the northern
route, reaching the hills by tonga from Gauhati to Shillong, the
capital of the Assam Province, and the residence of the Deputy Com-
missioner of the Khasia Hills, whose permission has to be obtained before
travelling in the outlying districts. There is a good road to Shillong,
but beyond that it is necessary to walk or ride, or to be carried by a
cooly in a thapa or chair. Coolies are obtained by applying to the
Superintendent of Police ; they are very strong, and carry loads of
70 lb. with ease over the roughest mountain country. There are good
NOTES.
263
dâk bungalows in the various villages, but all furniture and provisions
must be carried from Shillong, except to Cherrapunji.
The voyage up the Brahmaputra is interesting, but the vegetation
somewhat monotonous ; the river flows among broad sandy islands and
low banks covered with tall reeds, forest being only visible in the far
distance on the higher lands out of reach of the floods. The country
between Gauhati and Shillong is very interesting, but is said to be
very unhealthy, at least in the lower parts. The first portion of the
64 miles drive is like the foothill country near Mirigama, and then the
road passes into the jungle for many miles ; the general type of vege-
tation is not unlike the Nalanda forests, but there are several unfamiliar
forms. Linum trigynum is a beautiful sight along the banks at the
roadside. These foothill jungles are very rich, and contain manj’^
interesting forms, such as Dipteris Wallichii. Their unhealthiness may
be avoided by living in the hills above and making occasional descents
into them. At about 3,000 feet the scenery changes to open rolling
patana country with strips of forest, and above 4,600 feet the forests
in the more open valleys consist mainly of Pinus Khasiana, giving a
very northern look to the vegetation and scenery, which is increased
by the great number of familiar European genera to be seen among
the patana plants, far more even than on Horton Plains or at similar
elevations in the Sikkim Himalaya.
Shillong is near the culminating point of the hills, and is a pleasant
little sanitarium, commanding beautiful views of the Himalaya. The
country to the south, towards Cherrapunji, is a level-looking plateau
at first glance, but further acquaintance with it discloses the fact that
it is intersected in all directions by extremely abrupt and deep valleys,
which ultimately debouch into the plains of the great Surma Yalley
below Cherrapunji. The latter place stands on a level plain, which at
the edges (drops suddenly into splendid canon-like valieys about 4,000
feet deep, with numerous beautiful waterfalls falling over their
precipitous cliffs. The higher levels of the plateau are open grass
country, but the valleys are mostly wooded, especially on their northern
slopes, and the forests are both rich and interesting, while to the
botanist from further south there is a great fascination in the many
European-looking forms to be met with both in the jungles and on the
patanas. Cherrapunji itself has long been famous as the locality of
the greatest known annual rainfall. This, however, almost all falls
between April and October, and the winter is cool and dry. The plain
shows very clearly the effect of the enormous precipitation in its bleak
stony aspect and its very scanty flora, while the valleys below are
extremely rich. Probably, as Sir Joseph Hooker has pointed out,
there is no place on the earth where so large a flora is to be found in
so small an area, the Indian, Chinese, and Malayan floras meeting and
mingling here, while, though the elevation of the Khasia Hills is small,
the mountain flora is that of much higher elevations elsewhere. I
264
NOTES.
know of no better or more interesting headquarters for an ecological
worker than Cherra, and it has a good dâk bungalow and fine climate,
except in the rains, which may be avoided by going further north to
Shillong. It can be most easily reached from Calcutta by steamer
from Goalundo up the Surma Valley to Chattuk and native boat to
Companygunj, from whence it is a walk of about 25 miles with coolies.
Another centre from which interesting excursions may be made is
Mawphlang (Moflong) between Cherra and Shillong, but it is a bleak
cold place in itself.
My second excursion was into the Sikkim Himalaya in the Darjiling
and Kurseong Districts. Sir Joseph Hooker’s classical description
of the interesting flora and vegetation of this country gives all that
is necessary for preliminary information, but it may be worth while
to point out to European botanists that this region is now very
much opened up, excepting in Nepal. Darjiling is a large town,
within 22 hours of Calcutta by rail, and there are now resthouses
at easy stages throughout the Sikkim routes followed by Hooker
up to elevations of 12,000 feet. The flora has been well worked
from a purely taxonomic standpoint, but there is a vast field for
ecological work in this region ; botanists familiar with the Swiss or
Pyrenean floras should find it of particular interest. At the same
time the flora of the lower slopes, e.g.^ about Darjiling itself (7,500 feet),
is very like that of Horton Plains and other high levels in Ceylon, but
mingled, like the flora of the Khasia Hills, with many more northern
types. The whole formation and scenery of the Lower Himalaya, as
seen from the Darjiling Railway, is very closely similar to that of the
western side of the Ceylon mountains. From the foot of the hills to
Calcutta the line passes through cultivated country, which in the last
portion is very like the neighbourhood of Colombo, but evidently drier.
I take this opportunity of expressing my thanks to the various
Government Botanists of India, Curators of Botanic Gardens, and
other officials in the districts which I visited, for the hospitality and
kindness shown to me at every stage of my journeys, and for the help
rendered in arranging details of tours, &c.
J. C. WILLIS.
Opening of the Experiment Station at
Gangaroowa, Peradeniya.
In the article on the history of the Botanic Gardens with which this
volume opened, it was mentioned on page 12 that the chief desideratum
for a complete modern organization of the Department was the opening
of an experiment station or garden, where experiments could be
tried on the large scale with staple products or with new products not
NOTES.
265
yet staples. This station is now opened, the Government having
purchased at the beginning of this year the almost historical estate of
Gangaroowa, lying on the opposite side of the river to the Botanic
Gardens. The estate is of 550 acres, and of approximately horse-shoe
shape, forming the opposite bank of the Mahaweli-ganga up to the sky
line all round the north, east, and west sides of the Botanic Gardens,
while on the east side between the river and the hill there are about
200 acres of nearly level land with good soil. Visitors to the gardens will
remember that many of the most beautiful views are obtained at places
where the river curves, and that their beauty is due to the banks of the
river itself and to the wooded hills behind. All these hills are included
in the new experiment station, and are to be worked as forest reserves,
so that there will be no interference with the scenic beauty of the
neighbourhood. The experimental plots of economic plants will be laid
out on the lower and more level land of the estate, which at present is
mainly cultivated in cacao, pepper, cocoanuts, grass, arecas, and croton.
The estate is of historic interest in two ways. Round the bungalow
enclosure may be seen the remains of the earthworks of the former
Portuguese fort, the scene of a great defeat and slaughter of one of the
armies which invaded Kandy. About 1824 the estate was opened
in coffee, indigo, sugar, &c., by the then Governor, Sir Edward Barnes,
who was one of the very first pioneers in the planting enterprise,
which soon afterwards spread so rapidly over the central districts of
the Island.
The New Branch Garden at Nuwara Eliya.
The Government having decided to open a small branch garden for
experiments with cultivation of useful and ornamental plants in the
peculiar soil of Nuwara Eliya, a small site of about 5 acres has been set
apart for the purpose in the new park, and is now being brought into
cultivation. It will be worked, like the Badulla garden, as a branch of
Hakgala. The land lies at an elevation of about 6,200 feet, and is at
present mainly patana with scattered Rhododendrons and a small
piece of swamp.
The New Resthouse at Peradeniya.
This building, mentioned on page 21 of the present volume as in
progress, was opened to the public on 10th March, and is very
convenient for visitors to, and workers in, the gardens. It lies about
200 yards from the principal entrance to the Botanic Gardens, contains
dining and sitting-rooms, and four large bedrooms, and is fully furnished
(36)
266
NOTES.
with all necessaries. Visitors working in the Peradeniya laboratories
have a prior claim to some of the rooms, for which a charge of Re. 1 per
day is made. Meals are supplied according to arrangements to be
made with the resthouse-keeper.
Publications.
The Catalogue of the Library of the Peradeniya Gardens to the end
of November, 1900, prepared by Mrs. Willis, was issued in book form
in March, and may be obtained for Re. P50, or through Messrs. Dulau
& Co. for 2.S'. Visiting botanists will find it useful in showing them
what literature it is or is not necessary to bring with them for their work.
The following numbers of the first series of the “ Circulars ” have
lately been issued, completing the first volume ; —
22. — School, Bungalow, and Resthouse Gardens, and some
Hints on how to plant them. By J. C. Willis.
23. — Cacao Canker in Ceylon. By J. B. Carruthers.
24. — Camphor. By J. C. Willis and M. K. Bamber.
25. — Mosquitoes and Malaria. By E. E. Green.
Personal Ilotes.
On the 28th February, 1902, Mr. William de Alwis, Draughtsman of
the Department, retired, and his son, Mr. Alfred de Alwis, was appointed
in his place. The long connection of this family with the gardens,
referred to on page 6 of this volume, is thus likely to continue unbroken
for a considerable period.
William de Alwis Seneviratne, Mudaliyar, was appointed draughts-
man on 1st November, 1865, and has thus served for over thirty-six
years, under Drs. Thwaites and Trimen as well as under the present
Director. The library at Peradeniya contains many hundreds of
beautiful drawings of native and exotic plants made by him. and he has
also for many years been responsible for the upkeep of the herbarium,
in which work his extensive acquaintance with the native flora has
proved of much service to the institution. Among the published
examples of his drawings may be mentioned many of those in Trimen’s
Flora of Ceylon (the rest being by other members of his family) and the
drawings of butterflies, &c., in Moore’s Lepidoptera of Ceylon. The
honorary native rank of Muhandiram was bestowed upon Mr. De
Alwis in 1896, and the higher rank of Mudaliyar in 1901.
)
Studies in tlie Morphology and Ecology
of the Podostemaceæ of Ceylon and India.
TT is a matter of common knowledge with regard to
-L families that consist entirely of water plants, that the
morphology of the vegetative organs is usually complex,
that the flowers are of simple structure (whether this be
primitive or due to reduction), and that it is very difflcult
to determine the systematic position of the order. Well-
known instances are the families Potamogetonaceæ, Nym-
phæaceæ, Ceratophyllaceæ. To none do these remarks
apply with greater truth than to the Podostemaceæ, and
perhaps no family is as yet so imperfectly known. In
saying this I by no means forget or undervalue the splendid
work of Tulasne and Warming, but there is yet room for a
vast amount of research before our knowledge of this order
can be regarded as anything like complete or accurate. We
have, thanks to Tulasne, a good knowledge of the general
morphology of the flower and fruit, and thanks to Warming
of the morphology of the secondary shoots and thallus of
many genera, chiefly South American, but this is practically
all. The germination, the morphology of the primary axis,
the development of the mature plant from the seedling, and
[Annals of the Royal Botanic Gardens, Peradeniya, Vol. I., Pt. IV., September, 1902.]
BY
J. C. WILLIS.
( With Plates IV.-XXXVIIT.)
INTRODUCTORY,
(37)
Ik. 1
268 WILLIS : MORPHOLOGY OF THE PODOSTBMACEÆ
the general life-history are unknown even for the compara-
tively accessible forms, such as Podostemon Ceratophyllum,
and the systematic position of the order is a matter of great
differences of opinion, some writers putting it in one, some
in another place ; e.g.^ near to the Saxifragaceæ, Caryophyl-
laceæ, Lenti bo lariaceæ, Pistiaceæ, Lacistemaceæ, Elatinaceæ,
Nepenthaceæ, Sarraceniaceæ, Piperaceæ, and elsewhere.
The present paper is an attempt to clear up some of the
darkness surrounding our knowledge of this order, so far as
the Asiatic species are concerned. No one can be more
conscious than myself of the defects and gaps in the work,
but I hope that its publication may stimulate botanists who
have the opportunity to study these plants in their natural
habitats, where alone they can be properly dealt with.
As to many botanical students the order is nothing but a
name, it may be well at starting to mention that it consists
entirely of submerged water plants, living only in rapids
and waterfalls, firmly attached to the rocks on which they
grow ; that their vegetative organs frequently consist very
largely of thalli, which are very often of (phylogenetic)
root ” nature ; and that both in external form and internal
structure they simulate in the most remarkable way Algæ
(e.g.^ Fucus), liverworts, mosses, and lichens. A glance at
the plates at the end of this paper will show this.
My observations have as yet been almost entirely confined
to the Indian and Ceylonese species. By no means one of
the least troublesome parts of the work has been the purely
taxonomic labour of determining the limits of the genera
and species, which have hitherto been very inaccurately
defined. This work is as yet very incomplete, but so far as
possible I have worked out the Indian and Ceylon forms,
and the somewhat sweeping changes in the nomenclature
and taxonomy which I have found necessary have been
described in a preceding paper with the view of clearing
the ground for the present work.^ As there explained, I first
* Willis, A Revision of the Podostemaceæ of India and Ceylon, Ann.
Perad. I., 1902, p. 181.
OF CEYLON AND INDIA.
269
studied the Ceylon species, then visited Europe to study the
Indian forms, but found the material hitherto preserved to
be practically valueless, and visited India myself and col-
lected fresh material of all but two of the recorded forms,
and of several new ones. I am indebted for valuable
assistance to the various friends there mentioned. The
life-histories of the Indian species have still, however, to
be worked out upon the spot.
As there is no good account of this family in English, and
as Warming’s classical papers are written in Danish, it will
be well to commence with a short account of what is already
known about these plants. The chief literature extant is
given in the following list, to which reference will be made
by number only in the rest of the paper : —
List of Literature.
(1) Aublet, Hist, des PI. de la Guiane franc., L, 582, t. 233, 1775.
(2) Bâillon, Hist, des Plantes, IX., 1886, p. 256.
(3) Beddome, Anamallay Plants, Trans. Linn. Soc., xxv., 1865,
p. 223.
(4) Bentham & Hooker, Genera Plantarum, III., p. 105.
(5) Bischoff, in Flora, xxvii., 1844, p. 416, t. 1.
(6) Bongard, Mem. de I’Acad. de St. Pet., VI., ser. III., 69, 1834.
(7) Brown, B., Canoe and Camp Life in Brit. Guiana, London,
1876, p. 11.
(8) Cario, Anatom. Unt. von Tristicha hypnoides, Bot. Ztg., 1881,
p. 25.
(9) Du Petit Thouars, Gen. Nov. Madag., ii., 1806 (Hydrostachys).
(10) Endlicher, Genera Plantarum, p. 268.
(11) Engler, Podostemonaceæ africanæ, Engl. Bot. Jahrb., XX.,
p. 134, 1894.
(12) Gardner, Structure and AflBnities of P., Calc. Journ. Nat.
Hist., vii., 166.
(13) Goebel, in Pflanzenbiol. Schilderungen, I., 166, II., 331, 374.
(14) do. Organographie d. Pflanzen, pp. 212, &c.
(15) Griffith, in Asiatic Res., xix., p. 103.
(16) do. leones Plant. Asiat., t. 541-544.
(17) do. Notulæ ad PI. As., p. 376, &c.
(18) Hooker, Sir J. D., Flora of Brit. India, V., p. 61.
(19) Hooker, Sir W. J., leones Plant., t. 2,356, 2,357.
(20) do. Companion to the Bot. Mag., II., pi. xx., and p. 23.
(21) Humboldt & Bonpland, Plantæ Aequin., I., 39, t, 11, 1808.
270 WILLIS: MORPHOLOGY OF THE PODOSTEMACEÆ
(22) Im Thurn, Among the Indians of Guiana, Loudon, 1888, p. 105.
(23) Jussieu, A. de, in Delessert, leones Select., iii., 91-94
(Hydrostachys).
(24) Klotzsch, in Peters, Reise Mossamb. Bot., 506, t. 52
(Hydrostachys).
(25) Kohl, Kalksalze und Kieselsäure in d. Pfl., Marburg, 1889,
p. 249.
(26) Kurz, in Journ. As. Soc. Bengal, xlii., 1873, ii., 103. '
(27) Bindley, Vegetable Kingdom, p. 482.
(28) Martins & Zuccarini, Nov. Gen. et sp., I., 6, 1822.
(29) Michaux, Flora Bor.-Am., ii., 164, t. 44, 1803.
(30) Möller, Cladopus Nymani, Ann. Jard. Bot. Buitenz., xvi.,
1899, p. 115.
(31) Müller, Fr., in Nature, xix., 1879.
(32) Richard, in H. B. & K., Nova gen. et sp., I., 246, 1815.
(33) Spruce, Notes on Plants collected chiefl}^ in the Rio Uaupes
in 1852. Herb. Kew, quoted in Goebel, Pfl. Sch., ii., 335.
(34) Strasburger, Bau und Wachsth. d Zellhäute, 1882.
(35) Thwaites, Enum. PI. Zeylan., p. 222.
(36) Trimen, in Journ. Bot., xxiii., 1885, p. 173.
(37) do. Flora of Ceylon, iii., 415, and v., 386 (Willis).
(38) Tulasne, Podost. Synopsis Monographica, Ann. d. Sc. Nat.,
III. ser., t. xi., 1849, p. 87.
(39) do. Monographia Podostemacearum, Arch, du Mus.
d’Hist. Nat., vi., 1852.
(40) do. in Martii Flora Brasil., IV., pars 1, 1855, p. 229.
(41) Wächter, Weddellina squamulosa, in Flora, 83, 1897, p. 367.
(42) Warming, Familien Podostemaceæ, I. -VI., Kgl. Dansk.
Vidensk. Selsk. Skr. 6 Kaekke, ii., 1881, ii., 1882, iv., 1888,
vii., 1891, ix., 1899, xi., 1901.
(43) do. Podostemaceæ, in Engler & Prantl, Nat. Pflanzenfam.,
III., 2a, p. 1, 1890, and Nachtrag, p. 179.
(44) do. Note sur le genre Hydrostachys, Bull. Ac. roy.
danoise, 1890.
(45) do. Podostemaceæ, Flora of Koh Chang, Bot. Tidsskr, 1901.
(46) Weddell, Monograph in DC. Prodr., xvii., 1873, p. 39.
(47) do. A new African genus of P., Journ. Linn. Soc. Bot.,
XIV., p. 208, 1874.
(48) do. Sur les P. en gén. et leur distr. géog. en partie..
Bull. Soc. Bot. France, XIX,, 1872, p. 50.
(49) Wight, Icônes Fl. As., V., p. 31, t. 1,916-1,920.
(50) Willis, A Revision of the Podostemaceæ of India and
Ceylon, Ann. Perad., L, 1902, p. 181.
(51) do. Additions and Corrections to Ceylon Flora. Tri-
men, Flora of Ceylon, V., 1900, p. 386.
(52) do. in Nature, vol. 61, Nov. 1899, p. 68.
(53) Woodrow, in Rec. Bot. Survey India, I., p. 95, 1895.
OP CEYLON AND INDIA.
271
In earlier times these plants were commonly mistaken for
algæ or lichens or even mosses ; the resemblance is so great
in many cases, especially in the vegetative condition, that
such mistakes are quite excusable. The resemblance is also
indicated by the frequency of such specific names as
bryoides, fucoides, algæformis, lichenoides, &c. The first to
be described and made the type of a distinct family was
Mourera fluviatilis, found in 1775 by A.ublet (1) in French
Guiana. He describes it as growing upon the rocks at a
great cascade in the Sinemari river, where it is always sub-
merged, only the floriferous part coming above the water.
Little progress was made in the knowledge of these plants,
and comparatively few had been described, until the
publication in 1849 and 1852 of Tulasne’s monographs
(38, 39). He divides the order into 21 genera, describing 79
species, most of them for the first time. The second paper
is illustrated with beautiful figures, which show in a very
striking way the great variety of form which exists among
these plants. Tulasne’s descriptions and detailed analyses
are very accurate, much more so than those of some of his
successors. It will be noticed that in the preceding paper I
have returned almost completely to his definitions of the
generic and specific forms of the Indian Podostemaceæ. His
interpretations of the morphology, as is only to be expected
in a work of that period, are often misleading.
The systematic part of Tulasne’s work was continued by
Weddell (46), whose monograph has been the standard for
subsequent classifications of the order. His work, however,
at any rate so far as the Indian forms are concerned, is
inferior to that of Tulasne in accuracy and insight, and he
does not deal at all with the morphology or ecology, except
in a separate paper (48), where he treats the order with
reference to its geographical distribution, and in which he
calls attention to the limited distribution of the species as
compared with most water plants, noting especially the
curious fact that in the rivers which he studied in
South America the forms were different at each cataract
272 WILLIS : MORPHOLOGY OP THE PODOSTEMACEyE
passed in ascending the rivers. The same fact has been
noted by Goebel (13) in the rivers of British Guiana, and
my own work described in the preceding paper inclines me
to believe that the same is the case in the rivers of India
and Ceylon.
The latest and most complete account of the family has
been given by Warming in a beautiful series of monographs
(42), with which we shall now proceed to deal. In the first
he describes Podostemon Ceratophyllum, Michx., found in
the United States, and two South American species of
Mniopsis, showing that the mature plant consists of a thin
filamentous creeping “ root,” closely attached to the rock,
often by special organs which he terms haptera, exogenous
outgrowths of the root or stem, very common in the whole
family. They are sensitive to gravity and contact, grow
downwards to the rock and spread themselves out on it,
being fastened very firmly by the development of rhizoids
or by a gummy secretion. The creeping root may be termed
the tliallus ; it bears a more or less aborted root-cap, is green,
and takes part in the work of assimilation. From it, at
regular intervals and in acropetal succession, are developed
secondary shoots, arising endogenously in the tissues of the
root, and emerging and growing upwards through the water.
These shoots have a complex morphology of their own, and
ultimately bear the flowers, usually terminal on the branches,
which themselves arise in the under axils of dithecous leaves
(leaves with a stipular outgrowth on the under as well as on
the upper edges). During the vegetative season the whole
plant is submerged, but as the water falls the flowers
emerge.
In his second paper, Warming describes two new types
of form. The first occurs in the South American Castelna-
via princeps, which has a thallus composed of combined
shoot structures, united to form a flat creeping body bearing
leaves on the margins (I, c., pi. XIII.), and with the flowers
emerging from small cavities in the thallus during the dry
season when the water-level is low. The second occurs in
OP CEYLON AND INDIA.
273
the Ceylon species Dicræa elongata and D. algæformis (D.
stylosa), where the thallus is more or less dimorphic, a part
of it creeping on the rock and branching in the plane of the
rock, part of it consisting of long filamentous or ribbon-
like organs arising from the creeping parts and streaming
out freely in the water. These thalli are phylogenetically
of “ root ” nature. On the streaming thalli the secondary
shoots arise endogenously, as in Podostemon, but are only
single-flowered and very short.
In the third paper several species of Podostemon are
described;, which resemble P. Ceratophyllum in their
morphology. Some species of Apinagia and Ligea are also
described, in which there is a more complex shoot than in
Podostemon, which has many leaves, and drifts out in the
water, bearing at the usual season a complicated inflorescence.
The shoot and not the root is the principal growing and
assimilating part in these forms. Mourera aspera, which
has a sort of rhizome, bearing large leaves and complex
inflorescences, is then described.
The fourth paper deals first with the curious African genus
Hydrostachys, which Warming has since placed in a separate
family. The Abyssinian form Sphærothylax (Anastrophea,
Wedd.), which has both a flat lichen-like thallusand a tall erect
stem, is then described. The thallus is shown to be of “ root ”
nature like the thallus of Podostemon, and it bears small
endogenous flowering shoots all over its surface. This very
remarkable structure is repeated in Hydrobryum olivaceum,
a Ceylon species described in the same paper, only that here,
so far as Warming’s observations (made on spirit material
collected in the dry season) show, the plant has only the
thallus, and thus practically consists of a ‘‘root ’’alone ; this
is, as we shall see, not quite the case, but the tall primary
axis when seen in 1854 was mistaken fora distinct species, and
received the name Podostemon Gardneri. This very curious
plant is also described by Warming in an earlier part of the
same paper under the name of Dicræa apicata of Tulasne.
The full life-history is given below. Prof. Warming
274 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
also describes in this paper two species of Lawia. L.
foliosa is correctly described ; it has a flat creeping thallus
composed of combined shoot structures (as in some Castel-
navias), and bearing flowers. We shall see below that this is
also the case in L. zeylanica, but in the spirit material of this
Ceylon species, which was all that Prof. Warming had for
examination, and also in the specimens preserved in
herbaria, the growing points were not present, and so the
true nature of the thallus was misunderstood, and it was
described as a “ root ” structure like that of Hydrobryum
olivaceum.
In his fifth paper, Warming deals with Tristicha hyp-
noides, a form widely distributed over the world, and
having a peculiarly complex form of shoot arising from
a creeping root. He also describes several new species
of Mourera, Podostemon, &c., and creates a new genus,
Leiothylax.
Finally, in his sixth paper. Prof. Warming describes
several new species in detail, establishes several new genera
upon the Indian material, as already described in my first
paper, and deals with the morphology of Marathrnm. He
also describes the Indian species Griffithella Willisiana
(Hookeriana, var.) and Hydrobryum lichenoides (Polypleu-
rum acuminatum,Wmg.). This paper ends with a discussion
of the classification of the order, which has been already
mentioned in my preceding paper so far as it concerns the
Indian forms.
But little has been written by any other authors on the
general life-history or ecology of the Podostemaceæ. The
most readable account is that given by Goebel (I3j, who
describes the plants from his own observations in British
Guiana. He also points oat again, what was first noticed by
Weddell, that the range of distribution of the species is, as a
rule, very limited, and that even in one river there may be a
different set of species or forms at each successive cataract.
He describes the way in which the Podostemaceæ grow only
in rapids, flowering when the fall of the water level expose
OF CEYLON AND INDIA,
275
them to the air, and he deals with the morpholog3^ of the
forms found him ; this work was continued by his
pupil Wächter (41), who has described Weddellina in
detail. The only other recent paper dealing with any of .
the family in full detail is that of Möller (30) on the newly
discovered Javanese species Cladopus Nymani ; his account
is criticised and amended in Warming’s sixth paper already
quoted.
The special literature of the Asiatic forms has been already
dealt with in detail in my first paper, and from the morpho-
logical point of view mention need only be made -of the
work of Griffith (15-17), in which man^^ interesting observa-
tions may be found scattered in various places, and of
Gardner, who described man^" of the Ceylon and South
Indian forms for the first time, and also gave a very interest-
ing account of the order, with which he had also worked in
Brazil.
GENERAL CONDITIONS OF LIFE,
Hahitat. — So far as is yet known, all Podostemaceæ live in
rapidly moving water, usually in verj^ rapid or even broken
water, such as is found in the rapids and waterfalls of
mountain streams. They are only found attached to rocks
as a rule, but occasionally may be seen growing upon logs of
wood or other objects which have become firmly wedged in
the rocks. Each species appears to affect a particular class
of habitat, probably chiefiy determined by the speed, depth,
and degree of brokenness of the water, and the species do not
intermingle very much, as will be more fully described
under each. A description of the chief locality at which I
have studied these plants will thus serve in general as a
description of the usual habitat of the plants of the order.
All but one of the Ceylon species as 3^et known occur
abundantly at a place called Hakinda, on the Mahaweli-
ganga, at the extreme north-east boundary" of the Experi-
ment Station at Peradeniya, and about a mile from the
Botanic Garden, at an elevation of about 1,500 feet above
(38)
276 WILLIS: MORPHOLOGY OF THE PODOSTEMACEÆ
sea-level. The Mahaweli-ganga, as it flows round the Botanic
Gardens, is a swift smoothly-flowing stream about 80-100
yards wide during ordinary weather. It is subject, like other
. mountain streams, to sudden and heavy floods during rains,
and I have seen it rise 10-15 feet bi a day. The current is
very swift even in the driest weather, and the water is
always muddy, owing to the wash from paddy fields and tea
estates on the hills which it drains. In the driest weather
it has a pale straw colour, and in rains it becomes a thickly
turbid -Stream, the colour being that of strong coffee mixed
with a moderate amount of cream. The rise and fall of the
water-level is rapid and decided ; in former years, before the
mountains of the Central Province were so largely cleared
of their forests for planting purposes, this was probably much
less pronounced, and the water, too, was probably less
muddy, and so perhaps the plants submerged in it could live
at a greater depth.
At Getambe, just below the gardens, the river enters a
deep rocky gorge and becomes a furious torrent, resembling
the Cl}' de above Cora Linn, or the Wharfe near the Strid.
The stream is closely confined by hard steep gneiss rocks
full of potholes, and in many places is reduced to a fraction
of the width that it has at Peradeniya. This gorge continues
for about half a mile to Hakinda, where it turns a corner,
and at the same time the steep hillsides fall back a little from
the glen, and the river widens out into a broad basin, run-
ning among numerous small rocky islets. The current is
still very rapid, but has less depth and force, and it is there-
fore practicable to wade about in many places, an impossible
feat in the narrow gorge higher up. All along the rocky
part of the river just described the rocks are more or less
overgrown with Podostemaceæ, wherever they are covered
by water during ordinary weather. These plants grow only
on places Avhere the water is in constant motion, and never
in stagnant water. They even flourish on the rocks at the
sides of the waterfalls, with the furious current rushing
right over them. If the level of the water falls at any time
OP CEYLON AND INDIA.
277
those plants which become exposed to the air soon die, and
so we see a sort of tidemark traced upon the rocks by their
covering of Podostemaceæ. This mark indicates the average
level to which the water falls during their growing season,
above which it may temporarily rise during heavy rain, or
below which it may sink during exceptional dryness of the
weather.
Plate IV. shows a view of a portion of the rapids at
Hakinda, photographed in the dry season, with fruiting
Podostemaceæ on the rocks. These plants do not show very
distinctly in a photograph, owing to their dull gray colour,
which is not very distinct from that of the rock itself, but it
is not difficult to make out the band of exposed dead plants
forming a kind of tidemark on the rocks for some few feet
above the present water-level. At the places of most violent
rush of water the thalli are chiefly the flat closely attached
Lawias and Hydrobryums, while in the less violent parts the
Dicræas occur, and in the eddies Podostemon and Farmeria
are abundant.
The Rainfall and its Distribution. — Perhaps the most
important factor in the life-history and ecology is the rainfall
and its distribution throughout the year, as this factor
controls the level of the Avater, and hence the local distribu-
tion and the period of flowering of the plants. The following
table, taken from the meteorological report of the Surveyor-
General for 1898, gives the average rainfall for twenty-six
years at Kotmale (Gingran-oya),in the centre of the mountain
basin of the Mahaweli-ganga ; the distribution of this fall
is a fair index of the general state of the water-level of the
river at Peradeniya at the various times of the year : —
Month.
Inches.
Days.
Month.
Inches.
Days.
January ...
2-93
... 7
July
21*41
... 23
February,..
2-19
... 5
August
16*59
... 22
March
4-92
... 8
September ...
15*44
... 20
April
9*34
... 15
October
17*60
... 23
May
12*11
... 16
November ...
11*72
... 16
June
24*99
... 25
December ...
8*36
... 13
278 WILLIS : MOEPHOLOGY OF THE PODOSTEMACBÆ
Total 147*60 on 193 clays. Greatest fall in 24 hours 1T40
inches. [If a coefficient of about 25-30 be added to these
figures, they may fairly represent the average level, e.g.^
28-33 in January and 50-55 in June.]
It will be observed that there are two seasons in the year,
each of about six months’ duration, corresponding to the two
monsoons, beginning with much rain and gradually becoming
drier. The north-east monsoon begins about the middle of
October, and there is much rain until about Christmas, with
high water-level in the rivers. The rocks seen in the
photograph of Hakinda are very commonly all submerged at
this time, though the rapids are about 200 metres wide.
About the end of the year the rainfall slackens, and the dry
weather of this monsoon, the “ dry season ” of S.W. Ceylon,
sets in. The Podostemaceæ, lately deeply submerged, now
become graduallj^ exposed to the air as the level of the water
sinks, and as they become exposed, the flowers, which are
quite ready for anthesis, open, become fertilized, and quickly
ripen their fruits, shedding the seeds upon the dry rocks,
while the old plants usually die. In this condition they
remain until April, when the south-west monsoon begins,
with gentle winds and a good deal of rain. This first onset
of the monsoon is not violent, and is locally known as the
little monsoon,” but there is usually enough rain to raise
the water-level so high that all the seeds germinate. A fall
of water-level takes place in May, at which time it is usually
practicable to get at the plants once more and to search for
seedlings, and then about the end of May the “ big ” monsoon
sets in with violent wind and rain from the south-west,
causing the river to rise very high. During June and July
it usually remains high, but in August and September there
is usually less rain, and the level of the water sinks, of ten so
low as to expose some of the plants to the air. If this
happens the exposed parts of the plants usually die unless
quickly re-submerged, and they have not yet developed their
OF CEYLON AND INDIA.
279
Ilowers.'" The development of the flowers takes place
during the high water of the last two months of the year,
and by the middle of December the flowers are generally
completely formed, ready to open as soon as exposed to
the air.
The general life-history at Peradeniya is then briefly
this : the seeds are shed upon the rocks in January and
February, germinate in April ; the plants reach their full
vegetative development in September and October, and
develop their flowers subsequently ; the flowers are exposed
to the air by the fall of the water at the beginning of the
dry season, and the exposed plants wither and ultimately
die.
Hitherto the life-history of these plants has never been
thoroughly investigated, all descriptions having been written
from, material collected at the flowering season. During the
greater part of their life the plants are deeply submerged in
violent torrents of usually muddy water, and their study is
a matter of difficulty and even of danger ; during the height
of the monsoons it is practically impossible to get at them.
Fortunately, however,, there are often periods when the
water-level is lower for a short time, and by combining these
periods over several years I have been so fortunate as to be
able to work out the complete life-history for most of the
Ceylon forms, with the exception of the development of the
floral shoots towards the end of the year.
Podostemaceæ are found abundantly in all districts of
India and Burma, where there is rain in the south-west or
summer monsoon, but on the whole their vegetative season is
shorter than in Ceylon, and the more so the farther north we
go in each region, the rains beginning later and ending earlier
in the year. The following tables of rainfalls in various
localities where Podostemaceæ have been found (taken
* If submerged again within a very short time, rejuvenescence takes
place, new growing points being formed behind the withered portion.
280 WILLIS : MORPHOLOGY OF THE PODOSTBMACBÆ
from Blanford’s ‘‘ Climates and Weathers of India, Ceylon,
and Burma,” 1889) will illustrate this : —
Month.
Ootacamund.
Merkara.
Mahabaleshwar.
Lanauli.
Darjiling.
Cherrapunji. |
Shillong.
Moulmein. |
January
0-5
0-3
0*4
0*1
0*7
0*6
0*4
—
February
0*2
OT
0*1
0*1
1*3
2*6
0*8
0*1
March
1-2
IT
0*4
0*1
1*7
9*0
2*0
0*1
April
3-9
2-2
0-9
OT
5*3
29*6
3*7
3*0
May
6-2
6T
1*4
0*6
7*7
50*0
10*0
19*7
J une
6*0
25-8
47*3
28*1
28*4
110*0
17*0
38*4
July
5-6
42T
102*1
65*3
28*5
120*5
14*0
43*9
August
4-2
25-7
68*6
40*4
28*5
78*9
14*4
43*0
September ...
3-7
12-3
32*9
23*8
16-9
57*1
154
30*3
October
9*8
8-0
5*8
4*8
7*5
13*6
6*2
8*4
November ...
2-9
2-6
1*1
0*4
0*1
1*8
1*0
1*5
December ...
1-6
0-6
0*4
0*2
0*5
0*3
0*4
0*1
Total ...
45-8
126*8
261*4
164*0
127*1
474*0
85*3
188*5
Thus, in the Nilgiris the vegetative season is not much
shorter than in Ceylon, in the Merkara Ghats it is practically
over in November, and in the Bombay Ghats (Lanauli) in
October. In comparing these figures, the proportions rather
than the total fall should of course be taken. Thus, the
Bombay streams fall greatly in October, though the total
rainfall is still considerable. The Assam and Burma
districts have a rather longer wet season than Bombay.
The rivers in the higher levels of the Western Ghats of the
Bombay Presidency are represented in December, the period
at which I saw them, by long dry torrent beds, filled with
immense numbers of boulders and stones of all sizes, but
with occasional, usually steeper, stretches of firm rock. On
the latter, but not on the loose stones, however large, unless
firmly wedged in some immovable position, Podostemaceæ,
and especially Lawia, abound in the districts I examined.
During the rains, as may be imagined from the rainfall
tables, these streams are large and violent, but by December
OF CEYLON AND INDIA. '
281
many are completely dry, and in none is there more than a
mere trickle of water. In this the few Podostemaceæ that
have not yet fruited are found, their green or red thalli
shining brightly in the sun and forming a very pretty sight.
On the rocks above the water-level are thousands of fruits
borne on the dead withered thalli. Higher still, and this is
a circumstance of rare occurrence with the greater uni-
formity of level of the Ceylon rivers, may be seen dry plants,
which have been exposed so early in the season that they
have not formed any flowers, and at times even seedlings
may be found by any one who has seen them in the living
condition, and thus can recognize them. A good deal of
information was obtained from the study of the early stages
thus preserved. The rejuvenescence of the thallus, on the
other hand, which is so marked a feature of the Ceylon
forms, enabling them to overcome the difficulties caused by
changes in the water-level between the rainy seasons, is rarer
and of less importance in the Bombay districts, where the
rivers run quite dry.
At Darjiling the periodicity must evidently be very much
the same as in the Merkara district of the Western Ghats,
Only one species has been recorded from Sikkim, Hydro-
bryum Griflithii, found by the Rev. P. Decoly in a stream
near Kurseong. I visited this stream, and found it to be one
of the rivers of the foot-hills, rising at about 7,000 feet, and
thus not fed in summer by snowwater. No Podostemaceæ
have as yet been found in streams fed by melting snows.
In the Khasia hills of Assam the periodicity is very
similar. I spent part of December, 1901, among these hills.
The season had been fairly normal as regards rainfall, but
the water in the rivers was far less low than in the Bombay
Ghats, though low enough to have exposed most of the
Podostemaceæ. In general the rivers were like those of
Ceylon in this respect, and they evidently do not run dry
like the Bombay streams. Many of the plants showed a
certain amount of rejuvenescence where they were still
covered with water. Hooker, in his Himalayan Journals
282 WILLIS : MORPHOLOGY OF THE PODOSTBMACBÆ
(yoL II., p. 314, note), says under date September, 1850 : —
“ Podostemon grew on the stones at the bottom : it is a
remarkable water plant, resembling a liverwort in its mode of
growth. Several species occur at different elevations in the
Khasia, and appear only in autumn, when they often carpet
the bottoms of the streams with green. In spring and summer
no traces of them are seen ; and it is difficult to conceive what
becomes of the seeds in the interval, and how these which are
well knov/n, and have no apparent provision for the purpose,
attach themselves to the smooth rocks at the bottom of the
torrents. All the kinds flower and ripen their seeds under
water, the stamens and pistil being protected by the closed
flower from the wet. This genus does not inhabit the
Sikkim rivers, probably owing to the great change of
temperature to which these are subject.”
The specimens which were collected by Sir J. D. Hooker
at this time were in flower or fruit. Whether the statement
that the flowers are fertilized under water is always correct
or not should be settled upon the spot. My own visit was too
late in the year for this purpose, but I saw evidence leading
me to suppose that in Dicræa Wallichii there might have been
cleistogamic fertilization. In the Hydrobryums I did not see
any evidence of cleistogamy. Cleistogamic flowers occur in
Podostemon Barberi, so that there is no reason to doubt that
they may occur in others of the order, and indeed one is only
surprised that this is not a more common phenomenon.
We have thus seen that the depth of the water at the various
times of the year, itself regulated by the distribution of the
rainfall, is the immediate determinant of the phases of the
life-history, and we must now go on to consider the other
chief general factors in the life of these plants.
Temporary Variations in Depth of Water. — Not merely
does the grand period, as we may call it, of the water depth
determine the great outlines of the life-history, but the more
temporary variations in depth have a most important effect
on the ecology of these plants. In the slowly moving streams
affected by most water plants the changes of level are slow
OF CEYLON AND INDIA.
283
and slight, and there is at any rate little prospect of the
plants being actually exposed to the air, and that, too, on a
substratum of bare rock rather than of water-retaining mud.
In rapids and in mountain streams generally, however, the
water-level is liable to frequent changes, whether of sudden
rise or of fall sufficient to expose the plants to the sun and air
on the bare rocks, or in very shallow water. It is by no means
uncommon in August or September to find, at Hakinda, large
quantities of Podostemaceæ exposed on the rocks by the fall
of the water, and when exposed at this period the plants
cannot form flowers, and simply die if not re-submerged
within a short time. I am inclined to think that the gradual
adaptation of the Indian Podostemaceæ to meet this danger
has been one of the most marked features in their evolution.
It seems to show especially in two ways, in the enormous
capacity of rejuvenescence that the plants possess, and in the
great dwarfing of many of the most highly evolved forms ;
it probably shows also in the increase of number and
diminution of size of the secondary shoots, and in the
partial amphibiousness of the thalli. We shall return to
this at the end of the paper.
Ultimate Exposure. — Whatever may be the effect of the
temporary changes of water-level, the plant must ultimately
be completely exposed by the great fall of the water-level in
the dry season, and adaptation of the life-history to this fact
is an absolute necessity. This is clearly marked in the way
in which the flowering takes place only at this time, while
during the period of high water-level the plant is vegetative
only, storing up large reserves of material to enable it to
flower and ripen its seeds very quickly when the exposure
does come, because, though able to live for a short time above
water, the thalli are not usually capable of standing very
long exposure.
Aeratio7i.—A.noi\iQv important factor in the life-history is
probably the aeration of the water. Instead of the almost
stagnant water with a muddy bottom in which so many
(39)
284 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
water plants live, the Podostemaceæ have constantly and
rapidly moving water with a rocky bottom, and the water is
well aerated. The necessity which thus appears to exist for
many water plants of providing means of supplying oxygen
for respiration to their more deeply submerged parts is
absent in the case of the Podostemaceæ.
Light. — The depth of the water acts in another way upon
the dis tribution of the plants in the vertical direction
by regulating their light supply. The vertical range of the
plants in the Hakinda rapids is small, extending only over
about three or four feet, or less in many places. The lower
boundary is mainly determined, it may be well supposed, by
the amount of light there available to the plants. Possibly
the actual pressure due to the water may join in producing
the result, but most probably the light is the larger factor, and
both depend directly on the depth of the water. During
periods of high water the plants seem to tend on the whole
to grow upwards rather than downwards.
The depth varies considerably at different seasons, so that
the light supply also varies, and the effect is intensified by the
fact that the muddiness is usually greater in the deeper water.
The earliest stages of the life-history must go on in
comparative darkness (especially as there is less sunshine in
the wet weather), while in August, September, and early
October, the period of greatest vegetative activity, the plants
are exposed to fairly bright light, being at that period close
to the water surface. With the reserve materials stored up
during this period the development of the floral shoots
is ultimately carried out towards the end of the high water-
level and comparative darkness of the north-east monsoon,
and reserves are also provided for the ripening of the seeds.
In the more northern stations of the Western Ghats this
statement must be modified, for there is there no drier
period between two wet ones. The life-history in this
district has yet to be studied, but probably during the time
of greatest depth of water the plants are vegetative only,
while they store their reserves and develop their flowers
OF CEYLON AND INDIA.
•285
during the gradually decreasing water depth and increasing
light of the last months of the rains.
That the plants have need for a considerable amount
of light is shown by the fact that they are rarely to be found
in shady places where the water is not exposed to the
sunlight for at least some hours daily. They affect chiefly
open sunny rapids.
In this connection it may be mentioned here that nearly
all these plants have a great amount of anthocyan in the
surface cells, so that Avhen alive they have a red rather than
a green colour.
Light, as is well known, tends also to have a dwarfing
effect on plants. Progressive dwarfing is one of the marked
features of the evolution of the Podostemaceæ, but to what
extent, if at all, light has operated in this must be a matter for
future investigation.
Temperature. — With regard to temperature, there is little
to be said. The Podostemaceæ inhabit, with few exceptions,
the tropical zone, where the temperature of the water is very
uniform, and the few that live in more northern regions,
e.^., those of Assam and Ohio, carry out their life-history in
the summer months. The forms living in the low-country
of Ceylon or South India have a very constant water
temperature of about 80° F. (27° C.), and those of the hills in
these countries a similarly constant though lower temper-
ature, e.g.^ at Peradeniya of about 75° F. (24° C.), or at
Paikara in the Nilgiris of about 58° F. (15° C.). In the Khasia
Mountains of Assam the life-history is over by the time that
the coldest weather begins, and the coldest water in which
I found any species in that district in December, 1901, was
about 57° F. (14° 0.). The temperature is so uniform in most,
if not all, cases that it may probably be almost ignored
as a morphological factor, but it appears to regulate the
geographical distribution to a large extent. No Podoste-
maceæ have yet been found in really cold water, and even
in India and South America they are, so far as yet known,
absent from streams fed by melting snow.
286 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
Speed of Current. — The next factor to be considered is the
speed of the current of water in which the plants live.
Leaving out of account the effects of scour, to be considered
presently, there seems to be no speed of current in itself too
great for the plants to be able to live in it. Podostemaceæ
are frequently found in the very rapid currents flowing over
the waterfalls and in the swiftest parts of rapids. On the
other hand, it is not all the species that are able to live in the
very swift water. The swifter the current, the greater the
strain on the plants submerged in it, tending to drag them
down stream. As might be expected, all the Podostemaceæ
show very well adapted structures for resisting this strain.
The very swift waters are inhabited, as a rule, only by the
best adapted species, i.e., in general the most dwarfed and
lichen-like ; the edges of waterfalls and the violent rapids in
Ceylon and India are usually inhabited by the flat lichen-
like Lawias and Hydrobryums, while the larger Dicræas,
Griffithellas, and still more Tristicha, are found in places
with less swift and violent water currents. As is already
well known, the Podostemaceæ differ from other water
plants ill the absence of the intercellular spaces which are so
marked a feature in the latter, and this peculiarity is perhaps
partly connected with the fact that they require to lie flat
down upon the rocks to avoid the strain of the current as
much as possible. A fragment of one of these plants at
once sinks in water if thrown into it.
Scour. — An important factor 'in the local distribution at
any rate is the scour of the current, due to the quantity of
suspended matter contained in the water, whether at all times,
as in the Mahaweli-ganga, or only at seasons of high floods.
If a very large scour goes on by the rolling of stones, pebbles,
and gravel down stream, the Podostemaceæ are not able to
survive this action, and places of much scour are found to
be free of these plants, except in sheltered bays in the rocks.
On t£e other hand, the plants may live successfully for many
years in a particular locality, and may be exterminated or
nearly so by a high flood causing an unusual scour. To such
OP CEYLON AND INDIA.
287
a cause, I think, roust be attributed the fact that the river
below Toong in the Sikkim Himalaya, which in 1899
contained large quantities of Hydrobryum Griffithii, now
appears to contain none ; the great cyclonic storm, on the
occasion of the Darjiling landslips, must have caused an
immense scour. So also, though I searched the stream in
which Grifiath originally found the species just mentioned,
I was quite unable to find a single plant ; it seems to have
been exterminated there since 1835 by natural causes.
Substratum. — As might be expected, the common sub-
stratum upon which these plants are found is smooth water-
worn rock, but occasionally, when a log of wood or other
object has become firmly wedged among the rocks, plants
may also be found upon it. They are never found on
unstable substrata, with the exception of one or two species
which grow in slower water, and are found attached to
pebbles, which they usually fasten together by their creeping
thalli. This is often the case with the Indian Tristicha
ramosissima and the North American Podostemon Cerato-
phyllum. The actual composition of the rock seems to be a
matter of practical indifference, and it is very doubtful if
the plants absorb much or any food from the rock upon which
they grow, unless perhaps silica, with which they are often
very largely provided.
Deposit. — A considerable amount of silt is often deposited
on these plants, especially at times when the speed of the
current slackens and the water-level falls. It must affect
their assimilation, and at times is heavy enough to injure or
kill them.
Biological Factors. — Competition with other plants is a
factor of very slight importance indeed with all the more
modified Podostemaceæ ; they live in water so rapid that no
other flowering plant, and only occasionally any Cryptogam
other than a minute alga or two, is found there. The less
modified forms, such as Tristicha, oftep share their habitat to
some extent with aquatic mosses, and on rare occasions with a
288 WILLIS: MORPHOLOGY OP THE PODOSTEMACEÆ
fern or an Eriocaulon. In the dry season, when the water-
level fall slow enough to bring the Podostemaceæ to the surface,
they become covered with filamentous algæ, but by this time
it can matter but little to their success or life. These plants
escape from competition to a degree that is very rare in the
vegetable kingdom. They hardly even compete with one
another for position, as each species on the whole affects, as
already mentioned, its own »particular kind of locality, and
only to a small extent mixes with others. This absence of
competition is probably a very important factor in the
morphological features of these plants ; they have been
enabled to adapt themselves structurally, mainly in relation
to the physical factors of their environment, rather than to
the biological.
Animal life also seems to have but little effect on the
Podostemaceæ. The movement of the water is too swift to
alloAv them to be attacked by fish, and the only animals that
seem to have any effect are the larvæ of various water insects;
these feed greedily on the thalli, which are usually very rich
in starch. Indirectly, wading birds are probably of much
importance, because they walk about on the rocks with wet
feet in the dry weather, and the small sticky seeds must cling
to their feet, and thus probably be at times carried to suitable
places for growth in other localities.
Evidently, then, we have in the Podostemaceæ a group of
plants of singular morphological and ecological interest, and
one in which there is still a very large field of work open to
modern methods of research in these lines. The present
paper hardly does more than clear up the darkness that has
hitherto surrounded these organisms in regard to many inter-
esting features of their life-history and general morphology,
leaving a vast amount of work still to be done.
Not merely are these plants directly interesting in them-
selves, but they afford a group in which the connection
between the change in morphology and the changed ecology
can be well studied, as we have already a very fair knowledge
OF CEYLON AND INDIA.
•289
of the morphology of the ordinary submerged plants of quiet
waters. The most striking feature about their general mor-
phology is the remarkable similarity which they exhibit to
the lower forms of the vegetable kingdom, in particular the
mosses, liverworts, and algæ, and in the latter more especially
the algæ of moving water, such as the Fuci of the rocks on
the sea coast. This similarity may be only “accidental,” or
it may be due to the action of similar causes. The Podos-
temaceæ are evidently very plastic, and we shall see in the
Indian forms a series of plants of gradually increasing
plasticity, and find that on the whole as the plasticity
increases the plant becomes more and more like the lower
vegetable forms. Once the plasticity is well established, so
to speak, there seems little limit to it, the more so as these
plants have no competition with other living organisms to
modify their adaptation to the physical factors of their
environment.
We shall proceed to deal with the Indo-Ceylonese species
in order. As described in the preceding paper, they belong
to two widely separated groups, the Tristicheæ (Tristicha
and Lawia) and the Eupodosfcemeæ (the other six genera).
So far as possible the systematic order there given will
be followed, but for the sake of simplicity of treatment
certain genera and species will be described out of their
proper order. We shall deal first in a general way with
the genus, and with preceding literature relating to it,
and then with its species. The habitat of each will be first
described, then the dry season appearance, which hitherto
has been the only one observed by botanists, then the
germination if known, the development of the thallus, its
morphology, the appearance of the secondary shoots, the
life-history during the vegetative season, the development of
the floral shoots, the opening, structure, and fertilization of
the flowers, and lastly, the fruit, seed, and phenomena of
rejuvenescence, with any other special features of biological
interest. The genus as a whole will then be considered in
relation to the general conditions of life.
290 WILLIS : MOKPHOLOGY OF THE PODOSTBMACEÆ
As may easily be conceived, and as Warming has pointed
out, the anatomy of these plants, with their highly organized
dorsiventral thalli and their peculiar habitat and mode of
growth, is of great interest. To deal with it in detail in this
paper would, however, enormously increase its already con-
siderable bulk, and therefore I have dwelt only lightly on
the anatomical features, so far as they are necessary for the
treatment of the general morphology, reserving details for a
later paper. The same statement applies to developmental
features of the flowers, &c.
The genera will now be dealt with in the following order ;
Tristicha, Lawia, Podostemon, Dicræa, Griffithella, Willisia,
Hydrobryum, Farmeria.
TRISTIGHA.
[Du Pet. Th. ; Willis, Rev. Podost, Ind., Ann. Perad. I., p. 207.]
This almost cosmopolitan tropical genus is represented in
India by the single species T. ramosissima (Wight), Willis,
found in Travancore and Malabar, and to which I have given
a separate sub-genus Dalzellia. The other two species of
Tristicha are placed in a sub-genus Eutristicha»
Cario (8) has described the anatomy and to some extent
the morphology of the widespread form T. hypnoides. This
he collected among mosses in a stream in Guatemala, grow-
ing in little tufts. He commences by criticising the
accuracy of Tulasne’s figure (39), but, as Warming points out,
both authors are right in general ; we have here one of the
many cases of confusion that have been caused by the
complex morphology and the polymorphism of these plants.
Cario describes the creeping root- or rhizome-like structure
which is closely attached to the rock, and which bears the
tufted leafy shoots, on which latter the flowers arise. He
names this creeping organ the thallus, and regards it as an
organ sui generis — not a root, because it has no cap, and not
a shoot, because it has no leaves. He describes its branching
as endogenous. It is traversed by a single vascular strand,
OF CEYLON AND INDIA.
291
in which, are two xylem groups towards the upper side. The
delicate moss-like leafy shoots are borne endogenously on
the flanks of the thallus, and have thin simple leaves, one
cell thick, arranged in three ranks, one upper and tw'o at the
sides. They are occasionally branched, chiefly at the base.
The flowers are borne on lateral branches of these shoots,
subtended by a couple of larger leaves at the base of the stalk
of the flower. Warming (42, V.) gives a fuller and better
account of the morphology. He regards the thallus as root,
chiefly from analogy with similar organs in other Podoste-
maceæ. The secondary shoots may be divided into two
classes, those of limited, and those of unlimited, growth. The
former are assimilatory, not erect, dorsiventral in structure,
with tristichous leaves, and grow only to a comparatively
short length ; the leaves towards the outer end are usually
longer, and the last leaf seems to be terminal or nearly so,
and is on the upper side. The shoots of unlimited growth
are longer, more or less erect, with a more complex phyllo
taxy. Their branching is peculiar ; the branches are of
two kinds, and are usually arranged in pairs, one above the
other, alternately on the two sides of the main shoot. Of
the pair of branches at each node, the lower is a shoot of
limited growth, the upper is one of unlimited growth,
repeating the structure of the main shoot. These short
shoots in the case of T. ramosissima have been termed ramuli
in systematic descriptions, and the name is convenient for
use. The flowers are borne on the shoots of unlimited
growth.
In a later paper (42, YI.) Warming describes the various
existing herbarium specimens of Tristicha, and figures many
of them to show their branching and other morphological
features. He reduces them all to two species, T. hypnoides,
Spr., in America and Africa and T. alternifolia, Tub, in
Africa. The latter comm only has its ramuli forked into two.
The very curious shoot morphology here described is
found again in Weddellina, the only species of which, W.
squamulosa Tub, has been investigated by Goebel (13, p. 349)
(40)
292 WILLIS : MORPHOLOGY OF THE PODOSTEMACBÆ
and his pupil Wächter (41). I have also examined Prof.
Goebel’s material, which he has kindly presented to me, and
verified most of their observations. The fioral shoots here
arise from the root thallus independently of the vegetative
ones. The latter are branched in two ranks in one plane
like Tristicha hypnoides, and bear large numbers of curious
scale leaves, in many of whose axils are ramuli, while the
shoots also branch into shoots of unlimited growth. The
pairs of shoots seen in Tristicha do not appear here. The
leaves of the ramuli, instead of having nothing in their
axils, have what Goebel calls Kiemenbüschel, small shoots
bearing a few slender leaf -like organs. Similar tufts occur
on the surface of the thalli or leaves of some S. American
Podostemaceæ ; they appear to be assimilatory organs, but
whether they are homologous in all cases remains to be
discovered.
The Indian sub-genus Dalzellia is distinguished chiefly by
the three stamens of the flower, the connate leaves at the
base of the flower-stalk, and the absence of tristichy in the
ramuli. The last point is led up to by the form lately dis-
covered by Miss Lister in the Nile, T. alternifolia, var.
pulchella, Wmg., in which the lower part of the ramulas has
broad tristichous leaves, the upper narrow leaves irregularly
arranged.
In many respects the Indian species seems one of the
most primitive among the Podostemaceæ, though its anemo-
philous flowers may perhaps indicate it as reduced in some
respects ; we know too little, however, about the phytogeny
of the order to say whether it is descended from anemo-
philous ancestry or not. At any rate, T. ramosissima lives
in very much less violent water than most of the Podoste-
maceæ, and has, much more than most of them, the
appearance we are accustomed to look for in water plants ;
and in view of the great difficulties in tracing the
phylogeny of this family, it deserves most careful study in
detail.
OF CEYLON AND INDIA.
293
Trisficha ramosissima (Wight), Willis.
rPlates V.-IX. and XXXVIII.)
My first acquaintance with this species was from material
kindly collected for me by Mr. T. F. Bourdillon at Munda-
kayam in Travancore in July and December, 1898. Subse-
quently Mr, C. A. Barber collected a splendid supply of
material in S. Kanara, and I have myself obtained some in
the Anamalais. As I have already mentioned in a preceding
paper, most of the Indian Podostemaceæ differ in detail in
almost every locality in which they grow, and therefore in
describing the material I have examined I shall in general
mention its exact origin. The plant still needs detailed
investigation at different times of year on the spot where it
grows ; its germination phenomena in particular ought to
prove of considerable interest.
Hahitat.—^Ï1\LQ plant lives in slow-moving or somewhat
rapid water at elevations to 4,000 feet, forming tangled masses
of green weed. Mr. Bourdillon found it in waterfalls. It
is sometimes mixed with Griffithella Hookeriana.
Dry Season Appearance. — We shall begin with the dry
season appearance of the plants, as seen in Mr. Barber’s
Nos. 2,517, 2,518. The rocks above water are covered with a
muddy tangle of dead weed ; if we remove the loose parts
we find the rock thickly covered with creeping filamentous
rhizome-like thalli (PI. XXXVIII., fig. 6) running in all
directions, frequently branched, the branches usually at
right angles to the main thallus (PI. VII., 1). At frequent
intervals along the thalli are curious “ feet,” thin flat disc-
like organs springing from the sides of the thalli (Pi. VII., 1)
and spreading out upon the rock, to which they are fastened
very firmly. These discs are most commonly in pairs, one
on either side, and examination shows that the tall withered
shoots which we have removed spring from their upper
surfaces. These shoots can be seen to have a complex
branching ; they bear numerous solitary flowers (PI. VI.),
represented now by the stalked fruits, which all point one
294 WILLIS : MOEPHOLOGY OF THE PODOSTBMACEÆ
way, as the flowers emerged through the water. Frequently,
however, we find merely a very short one-flowered shoot
springing from the thallus, and in this case as a rule there
is no large foot. In specimens from other localities the feet
are not so markedly disc-like ; in material from the
Anamalais they are stouter, and more or lesslohed (PL VIL,
fig. 5). In Mr. Bourdillon’s material, again, and in some of
Mr. Barber’s, the foot is represented by a comparatively
long peg-like organ, which projects downwards from the
thallus, tapering almost to a point, but expanding at the tip
when it touches the rock, to which it becomes firmly attached
(PL VIL, 8, 9).
Mature Structure. — Closer examination of the thallus
must be made on living material. The dead plant sheds its
seeds upon the rocks, and germination takes place with the
onset of the rains. We have no knowledge of the early
stages in the life-history, or of the primary axis, though
we may assume that it gives rise to the creeping thalli.
Inspection of the tip of a thallus shows that it is almost
cylindrical, about 0*5 mm. thick, and that it has a well-
marked root cap (PL YII., fig. 2). Taking this fact together
with that of the endogenous development of the lateral
branches, we need have no hesitation in regarding the thallus
as of “ root” nature. The thallus may grow to a length of
many inches, and is more or less closely appressed to the
rock. When it touches it, it is fastened by root-hairs, which
spread out at their tips in the usual way (PL YII., figs. 3, 4)
and secrete a dark-coloured cement-like substance. A little
way back from the tip of the thallus the secondary leafy
shoots may be found in course of development (PL YII.,
fig. 5). In the Anamalai material figured the shoots appear
in approximate pairs, one on either side of the thallus,
breaking out endogenously from lateral expansions of it,
while at the same time haptera or fastening organs develop
from these expansions, underneath them. As these organs
will be constantly met with in subsequent descriptions and
occur in a great many Podostemaceæ, it will be as well to
OF CEYLON AND INDIA.
295
describe them here in a little detail, and reference may be also
made to Warming’s descriptions and figures. A hapteron
may appear at any part of stem or root which is near the
rock ; it appears as an exogenous cap-less outgrowth, grow-
ing by a meristem at the apex, and bending downwards
under the influence of gravity to reach the substratum, on
which it flattens itself out, often developes root-hairs, and
becomes very firmly attached by a kind of dark-coloured
cement which is secreted by the root-hairs, or perhaps some-
times by the surface cells. It very often branches exo-
genously, or is merely lobed. All cases may be seen in the
various figures given below and by Warming.
A hapteron may be looked upon physiologically as a
gigantic rhizoid, and acts in the same way as an ordinary
unicellular rhizoid, bending down to the substratum and
adhering toit, often becoming lobed or branched. The size
and form of the haptera in the species now under consider-
ation seem to depend largely upon whether the thallus is
touching the rock or not at the point where they form, and
also upon the strain caused by the water current. In the
Anamalai material, where the thallus touches the rock, the
hapteron grows into a more or less discoid foot ; in the
S. Kanara material, which grew in less rapid water, the disc,
as the figures show, is very regular and thinner, while in the
Mundakayam material collected by Mr. Bourdillon upon the
edges of waterfalls, where the strain is greater, the haptera
are larger, more lobed, and stretch further out upon the
rocks (PI. YII., fig. 10). The formation of a discoid hapteron
seems to be dependent on the thallus touching the rock ;
when, as frequently in the material from S. Kanara and
Mundakayam, the thallus is above the rock, a long tapering
peg-like hapteron is formed, which grows down to the rock
and spreads out upon it, while the secondary shoot appears
endogenously at the upper and outer side of the peg (PI. VII.,
fig. 9).
The formation of secondary shoots appears to be practi-
cally always accompanied by the simultaneous or previous
296 AVTLLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
development of haptera, as indeed we might expect on
mechanical grounds, and as the secondary shoot grows larger
the haptera tend to increase also both in size and number, to
provide the necessary power of adhesion to the rock. In
the S. Kanara material the hapteron ultimately forms a disc
in most cases, whilst in the material from the Anamalais
and Travancore it is usually branched, and this the more
as it increases in size. In some of the Travancore specimens
the hapterous feet at the bases of the full-grown secondary
shoots were as much as 2-3 cm. across (PI. VII., fig. 10),
forming a holdfast of great power. Part of this foot appears
to have been due to further development of haptera from
the base of the stem, a very common occurrence. The
endogenous shoot arises from the upper side, and is at first
more or less horizontal (PI. VII., figs. 5, 6, 7), but soon
bends upwards. It is always more or less flexible, and
sways with the movement of the water.
The thallus itself is of simple structure, with a simple
central vascular bundle, which seems very like that of
T. hypnoides as described by Cario, or that of Weddellina,
as described by Wächter (41^ ; I have not made a detailed
study of it. The cells near the edge contain much silica, so
that the thallus is very hard to cut in section, and also very
brittle. This development of silica is very characteristic
and abundant in very many of the Podostemaceæ, and has
been described by Cario (8), Kohl (25), Warming, and others
(and cf. Lawia, below). It is so plentiful in many thalli that
a razor is almost completely blunted in cutting a single
section. A physiological function in helping to make the
organs containing it more or less amphibious has been denied
to this silica development, but not, I think, altogether on
good evidence. That the thalli are more or less amphibious,
i.e., that unlike most water plants they are able to survive a
considerable period of exposure (and that, too, on naked
rocks), we shall have occasion to see below in the case of
several of the Ceylon species, and I am inclined to think
that the silica in the outer cells may have some function in
OP CEYLON AND INDIA.
297
making the inner tissues more retentive of their water, just
as is supposed to be the case in some of the Crassulaceæ,
e.g.^ Rochea falcata. The silica is by no means universally
present, nor in all parts of the plant ; it is confined mainly
to the thalli, which are the parts most able to produce new
growing points and resume their growth after a period of
exposure.
Passing on now to deal with the secondary shoots, which
in this species make up the bulk of the plant, do most of the
work of assimilation, and ultimately bear the flowers, we
shall consider first the most fully developed type of shoot,
and afterwards the reduced shoots mentioned above as some-
times found on the thalli.
The growing point emerges endogenously from the
thallus, as already mentioned. It is of an ordinary shoot type,
bearing leaves closely packed together. The phyllotaxy
is complicated, and I have not been able to make it out
satisfactorily. The leaves are of simple structure, very
delicate, not unlike the leaves of a moss ; they resemble those
to be described below for the ramuli, but are larger, and have
often a central portion more than one cell thick, like the
leaves of Lawia described below.
The most noticeable feature about the growing point is
the very early formation of shoots of the second order,
which grow very rapidly and cover up the main apex with
their leaves, thus rendering it difficult to dissect out the
principal growing point. These shoots are shoots of limited
growth, or, as we have called them above, ramuli ; they do
not repeat the structure of the main stem. The growing
point of a ramulus is figured in PI. VII., fig. 11. It has
a well-marked dermatogen layer, with inner meristematic
layers, and gives rise to a large number of leaves lower down,
these leaves being only one cell thick, four cells broad (in
the S. Kanara material), and exceedingly delicate. The tip
of one of these leaves is shown in PI. VII., fig. 13, and
the basal part in optical section in fig. 14. The basal cells
298 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
have rather thicker walls. The leaves contain chlorophyll,
and form the chief assimilating organs of the plant. In the
material from the Anamalais and Travancore the ramuli
have broader leaves (PL YII., cf. figs. 15, 16). The ramnlus
grows only for a short time, and soon reaches its full length.
The general appearance of a plant covered with these ramuli
is well shown in PI. V. The ramuli near the tip of the
stem far overtop the main axis. Later in the life of the
stem the ramuli either drop off altogether, or lose their
leaves ; the latter case is very evident in the right-hand part
of PI. YI.
It is very difficult, as Warming has already pointed out,
to determine the exact relationship of position of the ramuli
to the leaves borne on the main stem, and I have not yet
succeeded in satisfying myself on this point. When very
young, the ramulus seems often to lie exactly in the axil of
a leaf, but when older and expanded it is very commonly
not so arranged, but frequently has a leaf a little to one side
of itself, as in PI. YII., fig. 12. There is not as a rule a
ramulus for every leaf. The axis of the ramulus contains a
very slender central vascular strand.
A little way back from the growing point, at the part
where the ramuli have reached their full development or
nearly so, buds may be seen in the upper angles between
them and the main stem (PI. YII., fig. 15) ; these buds give
rise to shoots of unlimited growth, which repeat the struc-
ture of the main axis. Commonly many of these buds
remain more or less dormant, until the flowering period.
Thus at every node there may be two shoots, an upper long
one of unlimited growth repeating the structure of the main
axis, and a lower ramulus. By this time the leaves of the
main axis have usually fallen away, leaving no trace, so that
when, as often happens, the ramulus has also lost its leaves
one is liable to mistake the two shoots for a leaf with shoot
in its axil. In Plate Y., and still better in the left-hand part
of PI. YI., this arrangement of the shoots can be very
clearly seen, especially with the aid of a lens.
OP CEYLON AND INDIA.
299
These secondary shoots, thus constructed, may grow to a
length of 30-40 cm. and branch freely, and as 20-50 may
form on one thallus, and the thalli often branch very freely,
while many plants may grow intermingled on a single piece
of rock, the result is a great tangle of vegetation. Plate Y.
represents a single plant, not quite complete.
The internal structure of the secondary axis in the vege-
tative condition is simple. A cross section near the apex
(PI. YII., fig. 17) shows that the stem is cylindrical, with
a well-marked epidermis, parenchymatous cortex, and central
vascular strand. Further down (fig. 18) the stem is thicker,
and two layers of collenchymatous cortex have been deve-
loped, one at the margin and one round the vascular tissue.
The bundle itself is central, and of very simple construction,
composed of phloem-like tissue containing sieve'tubes, mixed
with an irregular mass of parenchyma ; there are no pitted or
lignified xylem elements so far as I have observed. Across
section of the bundle is shown in PI. YIII., fig. 1. A certain
tendency to grouping, where recent divisions have occurred,
is seen in the tissue, but more detailed study is required to
make out the exact anatomical relationships of all the tissues,
and such is not necessary for this paper. The chief feature of
general interest in the section is the entire absence of inter-
cellular spaces of the kind one is accustomed to look for in
water plants ; this feature is general to the whole order, as has
frequently been pointed out. Its meaning will be discussed
later.
Further down in the stem thei’e is often to be seen a small
amount of somewhat irregular cell division going on in the
outer parts of the vascular strand, but there does not seem to
be any regular cambium layer formed. The cells of the
cortex enlarge considerably in the older parts of the stem,
but 1 have not seen any growth of the stem in thickness by
actual division of these cells as described by Wächter (41) in
Weddellina.
We have now brought our consideration of the vegetative
growth and structure of the plant to the full-grown stage
(41)
300 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
represented in Plates V. and VI. (left side). The former
especially gives an excellent representation of the habit of
the mature plant. I owe its success to my friend Mr. Barber,
who prepared the herbarium specimen from which the
photograph is taken, by cleaning the plant of algæ and mud
and then floating it out upon paper under water, as in the
method usually employed with delicate algæ. The creeping
thallus is clearly shown with the numerous secondary shoots
springing from it.
On some of the Travancore material shoots were found of
the kind represented in PI. VIII., figs. 2, 3. Here the secon-
dary axis had given rise to branches which grew out only
for a short distance and then seemed to check in their growth,
forming short spurs with a number of ramuli.
Towards the end of the wet season the plant developes
its flowers. As a rule, the tip of the shoot does not become
floriferous, but most of the lower portion does. What appears
to occur (further investigation is required through the whole
process, on living plants) is that in the axils of many ramuli
or former ramuli, the buds hitherto dormant form floral
shoots. Almost every axis other than near the tip of the
plant seems to form a flower. The flower is terminal on the
axis bearing it, and its pedicel is surrounded by the upper-
most leaves of the axis, the innermost of which are more or
less connate into a shallow cup round the base of the pedicel
(PI. IX., figs. 2, 3). These leaves are a little larger than most
of the other leaves of the plant. In the axils of two, or
sometimes three, of the lowest leaves on the short flowering
shoot ramuli are developed, so that the flower is almost always
accompanied by a couple of ramuli, as described by Wight.
Some of these ramuli can still be seen at the bases of the
shoots (now in fruit) figured in the right-hand half of PI. VI.
In this way a flower is formed at nearly every node in the
plant, but the buds on the outer parts of the branches of the
shoots usually remain in a very rudimentary or abortive
condition. It will be noticed that at the base of nearly
every branch, in the place formerly occupied by theramulus
OP CEYLON AND INDIA.
301
on the main stem, there is a floral shoot. This seems to
develop from the same axis that once gave the ramulus, but
of this I cannot be certain from the material at my disposal.
Instead of being, as the ramuli were, exactly under the
branches of unlimited growth, the flower shoots are usually
displaced upwards towards one side.
As Plate VI. shows, the floral shoots are very short, con-
sisting simply of the very short axis with leaves, ramuli,
and cupule, with the flower stalk emerging from the latter.
The flowers certainly open in the air before the water has
fallen so low as to strand the shoots upon the rocks, but
whether the pedicel lengthens while still under water so as
to bring the flower above the surface, as in many ordinary
water plants, or whether it remains of constant length till
the flower is exposed to the air, I cannot certainly say ; the
former is more probable from analogy. All the floral shoots
bend upwards, evidently being very sensitive to the stimulus
of light or of gravity at this period.
Before proceeding with the consideration of the flowers
we must go back to deal with the shoots mentioned as often
found upon the thalli, which are not fully developed like
those we have considered. Very often there may be seen a
couple of ramuli apparently springing from the thallus ; on
closer examination these are found to spring from a short
secondary axis which has not lengthened in the ordinary
way. A not infrequent type of shoot to be found on the
thallus is that shown in PI. VIII., figs. 4, 5. A short stout
axis, bearing leaves which persist somewhat longer than is
usual with, leaves on the main axis, springs from the thallus
to a height of 1-2 cm., with a stout hapterousfoot. It some-
times branches once or twice, but very commonly is quite
simple, usually bearing one or two ramuli, and ending in a
flower, at the base of whose stalk is a cupule, as usual.
Sometimes, as shown at x in fig. 5, the floral shoot is even more
simple, and is like one of the shoots already described which
arise at the nodes of the main axis. This type of shoot is
very frequent on the thalli of the S. Kanara material, which
302 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
Mr Barber describes in his notes as giving the impression
that the rock itself was covered with flowers. Several such
shoots are shown in PL XXXVIII. (right-hand rock), but are
difficult to distinguish.
The flower stands on a short pedicel, which lengthens
rapidly during and after anthesis. The structure of this
pedicel is interesting, and it will save repetition to describe it
here, as the same features occur in most of the subsequent
species to be described. At the time when the flower opens
the pedicel is short and stout, and shows the central vascular
strand quite clearly through the pellucid cortical tissue. As
seen in section (PI. VIII., fig. 1) it has an epidermis, thin
walled cortex, and a central vascular strand with more or
less pith cavity in the middle. After flowering the pedicel
lengthens, and at the same time the small-celled tissue
surrounding the vascular tissue becomes thick walled and
lignified (PI. VIII., fig. 8, lig.), the vascular bundle itself
remaining as before. Finally, when the fruit is nearly ripe,
the outer cortex falls away altogether (fig. 9), leaving the
fruit standing on a long filamentous stalk, which is very
much thinner than the original stalk of the flower. The
vascular bundle itself is shown in PI. IX., fig. 1. In the middle
is the intercellular space, with several annular vessels in
it (somewhat as in Maize), and round this is some phloem-
like tissue and the outer fibrous band, which ends off sharply
against the parenchymatous cortex.
The flower itself is simple (PL VIII., fig. 6), and requires
only a brief description. The perianth is membranous,
united to about | of its height, and finally marcescent. There
are three stamens alternating with the segments of the
perianth, with long flexible filaments and large anthers with
loose powdery pollen. The ovary is superior, with three loculi
and three long hairy stigmas. The flower is anemophilous,
and probably largely self -pollinated ; most of the flowers
seem to set a full complement of seed. In transverse section
the ovary wall and structure are like those of Lawia, with
OF CEYLON AND INDIA.
303
ribs in corresponding places, and after fertilization a fruit of
exactly similar type is produced {cf. below).
Rejuvenescence. — To what extent rejuvenescence, Le., the
formation of new growing points if the old ones are damaged,
and the consequent resumption of growth of the thallus and
formation of new secondary shoots, takes place in this
species, I do not know, but there is evidence upon the
material in my hands to show that it does take place, and
probably it is common enough, as in the Ceylon forms.
Before going on to the next genus, we must briefly cou“
sider the morphology of Tristicha in connection with its
ecology, or in relation to the general conditions of life. All
three species seem to be very similar both in morphology
and ecology. Comparing these plants with most other water
plants, it is evident that they are as highly modified to suit
a submerged existence as any others, and that they may
well be compared with the members of such families as the
Ceratophyllaceæ. They show most of the characters one is
accustomed to look for in submerged plants, e.g., much
vegetative growth, frequent branching, great vegetative
reproduction, absence of stomata, reduction of vascular
tissue, especially of the water-carrying elements, central
vascular bundles, delicate leaves, chlorophyll in the epi-
dermis, absence of palisade tissue, &c. On the other hand,
they show a few marked features which are not usually
present in such plants.
To take these in order, the thallus, speaking in the ordi=
nary sense, is evidently morphologically a creeping root,
such as in other water plants and even land plants is not
uncommon, where the primary root owing to the mode of
growth in relation to the substratum is unable to develop
in the ordinary way, and is replaced by adventitious roots
laterally developed. This development is carried to an
extreme in the plants now under consideration, as the root
is used as the means of large vegetative multiplication of
304 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
the shoots, instead of being, as sometimes is the case with
such adventitious development of lateral roots, intended
mainly for anchorage or absorption. There is no great
assimilatory function performed in the thallus in this
genus, bu,t we get a distinct stage on the way to such cases as
Dicræa or Hydrobryum. Anatomically, almost more than
morphologically, the root shows great dorsiventrality.
In the haptera, we meet with organs for which there
appears to be no precedent in other water plants. Their
function is evidently to act as holdfasts, and they may
therefore be regarded as adaptations to the peculiar mode of
life, except that, so far as we know, they are not hereditary
in the strict sense, but are developed apparently in response
to a direct stimulus, at any point where they are required.
Their size and form also seem to be directly dependent on
the mechanical conditions of life.
The secondary shoots are endogenously developed on the
root-thalli, and here we meet with one of the most marked
general features of the whole order. Such formation of
secondary shoots upon roots is not entirely unknown else-
where, but it is not carried to the extreme pitch of frequency,
regularity, and complexity that is found in this family.
In the absence of other families of plants living under
similar conditions of life, or showing this peculiar feature
under different conditions, it is difficult to decide whether
this development of numerous secondary shoots is really an
adaptation to the conditions of life or not, but I am inclined
to think that it is. In this place, however, it is sufficient to
point out its advantages, in that it reduces the risk of ex-
termination when a temporary fall of water occurs other
than at the regular flowering season, which, if there were
only one axis, would be very great.
The secondary shoots are large and complex, but have an
evident aquatic habit and structure ; the vascular tissue is
reduced and central, and there is great development of
green assimilating tissue by means of the formation of the
ramuli.
OF CEYLON AND INDIA.
305
With regard to the absence of intercellular spaces, Goebel
(13, p. 354) has already given what I am inclined to think is
the true explanation or near to it, that the Podostemaceæ,
living in well aerated water, do not need, like plants whose
lower parts are in mud or in stagnant water, large intercel-
lular spaces to serve as oxygen carriers to the lower parts for
their respiration. I do not think that the very common
explanation of the large spaces, that they serve to float the
plants up in the water, is a sufflcient one, though this may
be part of their function. Tristicha floats freely enough
with the current, though it has no such spaces, and is
actually, like all the other Podostemaceæ that I have
examined, so heavy that it sinks in still water. In general
the development of large spaces seems to go with absence of
oxygen in the lower part of the medium surrounding the
plants, and with still water, both of which conditions are
absent in the case of Tristicha.
The great number of flowers produced, and their develop-
ment all at one season, is another noteworthy feature as
compared with plants of still waters ; the production of the
flowers at the time when the water leaves the plant stranded
may be compared with the development of the sexual organs
in Riccia fluitans when it is stranded by the drying of the
pool in which it has been floating.
Lastly, the power of rejuvenescence of the thallus, which
is well-marked here, and much more so in species to be
considered later, is evidently a very great advantage to a
plant exposed to the risks attending the temporary fall of
water-level that are liable to occur.
In Tristicha and Weddellina, then, we seem to have plants
still very like ordinary water plants, but with certain new
features correlated with their new conditions of life, such as
the development of creeping roots and haptera as fastening
organs, the multiplication of the shoots by the exaggeration
of the not elsewhere unknown phenomenon of development
of shoots from the roots, great powers of rejuvenescence.
306 WILLIS : MOEPHOLOGY OF THE PODOSTEMACEÆ
development of flowers all at one season, and absence of
intercellular spaces ; the plant is finally destroyed by ex-
posure to the air, leaving only the seeds and perhaps a few
submerged thalli to tide it over the dry season.
LA WIÄ.
[(TERNIOLA), Tul. ; Willis, Rev. Podost. Ind., Ann. Perad. L, p. 209.]
This peculiar little genus is confined to Ceylon and
Western India, where it is extremely common in suitable
localities. In Hooker’s flora seven species are recognized on
Weddell’s authority ; of these, I have transferred L.
ramosissima to the genus Tristicha, while of the remaining
six, three proved to be completely identical with forms
described under other names, thus reducing the genus to
Tulasne’s original three species, zeylanica, pulchella, and
iongipes. I hardly consider these latter two separable, as
the length of leaves and pedicels are, as we shall see below,
extremely variable characters ; I have therefore united
them, and have also united the Indian forms to the Ceylon
ones to form the one species L. zeylanica, of which there
are many varieties and sub-varieties. As a temporary
measure, pending detailed knowledge of material from
many sources, I have divided the species into two Ceylon
and two Indian varieties, Gardneriana, Parkiniana, malaba-
rica, and konkanica, respectively.
The morphology of the vegetative organs has not been
accurately described in this genus. The accounts of
Gardner (12), who first discovered it, and of Tulasne, though
accurate as far as they go, are very little more than descrip-
tions of the flowers. The thallus adheres so closely to the
rock that it is very difficult to detach more than minute
fragments for herbarium specimens, and as the flowers are
usually most numerous on the older parts, it so happens
that practically none of the existing specimens show the
growing points or the mode of growth of the thallus.
Tulasne, indeed, noticed that not all the leaves were in the
rosettes that are the most conspicuous feature of the upper
OF CEYLON AND INDIA.
307
surface, and Trimen describes the flabelliform apices,
though he did not notice their leaves. Warming’s account
of this species (42, IV.), though correct enough in its descrip-
tion of the facts observed, suffers from this lack of complete
material ; he had only the central flowering part of a thallus
to deal with. Upon this there are no leaves borne directly,
and hence Warming writes “le thalle ne porte pas directe-
ment de feuilles, ce qui indique que c’est une racine étalée
en forme de thalle crustacé.” Eoots of this form are not
uncommon in the order, e.g.^ in Hydrobryum, and there is
almost literally nothing to distinguish the central part of
the thallus of Lawia from a “ root ” thallus. Warming’s
account does not deal with the life-history or with the
growth and development of the thallus. In the same paper
he also deals with a specimen of L. longipes or foliosa
collected at Khandala in the Bhor Ghat by Goebel (13), who
has himself also described it. Here the incompleteness of
the description is in the other direction ; only the growing
point was examined, and not the older parts of the plant,
and it was at once evident that the thallus was of stem
nature. In reality both thalli are to all intents alike, but the
difference between the apical and the central parts is so
great that it is impossible to understand the morphology of
the plant without seeing a complete specimen. In his
sixth paper Prof. Warming has suggested that the L. zey-
lanica forms should be generically separated from the L.
foliosa forms on the ground of their thallus morphology,
but it will be seen below that this proposal rests on incom-
plete knowledge of the entire plants.
Lawia ^eySaiiica, TuL
(Plates IX.-XIII.)
My studies of this species have been made at Hakinda
and other places in Ceylon, and at Igatpuri and Khandala in
the Bombay Ghats, while I am also indebted to Mr. 0. A.
Barber for much valuable material from many places in
(42)
308 WILLIS : MORPHOLOGY OP THE PODOSTEMACEÆ
South India. The detailed description given below refers
in general to the common Hakinda form, Gardner’s original
type of the species, which I have described as L. zeylanica
Gardneriana, but the points of difference and of interest
found in the other forms are also described.
Habitat. — The Ceylon forms of this species grow at
elevations from 1,000 to 2,000 feet. It is found on smooth
rocks in places where, during the vegetative season at least,
the flow of water is very swift, but as a rule not in places
where it is also very much broken. Most often the plants
cover the rock to the exclusion of other species, but they
are frequently mixed with Hydrobryum olivaceum, less
often with Farmeria metzgerioides, and occasionally (usually
at junctions of eddies with the main stream) with Podos-
temon subulatus or Dicræa stylosa, var. fucoides, all of w^hich
are also very dwarf forms, or forms which lie very low
upon the rocks. The habitats of the Indian forms are, so
far as I can judge from their dry season appearance, very
similar. I have found them mixed in many places with
Hydrobryum lichenoides, and some of Mr. Barber’s
specimens are mixed with Griffithella Hookeriana.
Dyy Season Appearance. — The plant is exposed to the air
early in the year (e.^., I found it in flower at Hakinda on
9th January, 1898, 8th January, 1899, and 17th December,
1899), flowers as soon as exposed, and very quickly ripens
and sheds its seeds. It is at this period that all the existing
herbarium material has been collected. Large expanses of
dry smooth rock, often many yards across, may be seen
covered with a brown coating of thalli, thickly studded with
ripe fruits, while close to the water’s edge may be seen
flowers in all stages from opening bods to nearly ripe fruits.
Several pieces of rock are shown in PL X. As a common
rule there is little form or structure to be seen in a hasty
glance, but often such a specimen as the large one there
figured may be found, in which it is evident that the thallus
is branched on a definite system, and is not a continuous
OF CEYLON AND INDIA.
309
sheet, as it seems to be in the other figures. Closer exami-
nation shows that it is really branched in an irregular
manner, so that any small portion of it consists of ribbon-
shaped strips, rarely more than 1 cm. wide. The actual
growdng apices, which at this season may still often be found
just below the water surface, are more or less fan-shaped.
Closer inspection with a lens reveals the rosettes of small
linear leaves which are scattered among the floral shoots and
usually arranged in more or less regular longitudinal rows
along the strips of thallus. The floral shoots themselves
will be found to be at the edges and tips of the thallus
branches ; each has one flower which emerges on a short
pedicel from a small bristly cupule. In herbarium specimens
the leaf -rosettes and even the cupules are often represented by
scars or little pits. In the large specimen in PI. X. the
cupules can be clearly seen, the fruit stalks having fallen out.
Germination and Life History .—T'h.Q seeds are shed upon
the rocks and upon the dead surface of the old thallus,
where they remain until the water rises to cover them. The
epidermal layer of cells of the testa swells up when wetted,
like that of the seeds of Linum, &c. This has often been
described as an adaptation to fasten the seeds to the rocks
for germination. It is true that when it dries after wetting
the seed is very firmly fastened to the rock, but the next
wetting washes the seed off once more, and it does not seem
probable that this mucilaginous coat can have anything to do
with preparation for germination ; it seems rather to enable
the seeds to be carried away by the feet of wading birds.
These may often be seen walking on the thalli in the dry
weather, and some of the minute sticky seeds must be
carried away by them. In fact, this is almost the only
conceivable method in which these plants can be carried
from one river to another. The number of seeds produced
is very large, each flower, as a rule, setting a full complement
of about 200-250, and the whole rock being covered with
fruits, but of all this multitude very few ever come to
310 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
anything. Many are washed away by the water, and only
those which are retained by a crack or crevice in the rock or
the old thallus seem to have any chance of germinating where
they fall. Once carried off by the stream, there is little
chance of a seed reaching a place where it can grow. It must
have a rocky substratum, in a rapid stream of water, and it has
no adaptation of its own, other than small size, to enable it
to retain a position in such a place.
Germination, as a rule, occurs when the water rises with
the rains of the little monsoon ; I have found seedlings in
the period of low water in May between the little and big
monsoons. In India the germination must evidently be later
the further north the locality.
The early stages are extremely hard to find in this species
and I have only been successful in getting a very few. It is
no easy matter in working with these plants to determine
which seedlings belong to which species, and two consecutive
years of observation were needed to settle these points.
The seedlings are excessively small, and can only be obtained
when the water-level is low, an unusual occurrence at the time
of year when germination occurs, except just for a short
period between the two monsoons. I have frequently
obtained them, however, by groping in deep water for
capsules which I knew to be those of the species sought;
every now and then a seed will be found to have germinated
inside the persistent capsule-valve. The seed is exalbumi-
nous ; the hypocotyl is short and stout, and the cotyledons
are flat when expanded, but crumpled in the embryo (PI. IX.,
4). When wetted the seed presently swells and bursts the
testa, the hypocotyl bends down at once to the rock, and
immediately becomes attached to it by a copious development
of rhizoids from the superficial cells (IX., 4); these organs are
unicellular, and flatten themselves out at the tips against the
rock, just like those of the thallus of Tristicha. At this
stage the seedling is extremely small, about 0.5 mm. high.
The base of the hypocotyl now begins to expand and forms
a larger surface of attachment, fastened by more rhizoids.
OF CEYLON AND INDIA.
311
No haptera were ever observed to form. The next stages are
shown in PI. IX., figs. 5, 6. Here the hypocotyl has formed
a kind of tuber, on the upper surface of which a few leaves
have developed. There is no elongation of the epicotyl in
the ordinary way. The order in which the early leaves
appear is very difficult to make out ; only a small quantity of
material in very few stages was at my disposal, and the
plants are exceedingly small and delicate. I am inclined to
think that the order is no very definite one ; sometimes, and
perhaps most often, the first two leaves are more or less
at right angles to the cotyledons, which are themselves
either opposite or approximately so, but I have found cases
in which the first two leaves were approximately parallel to
the cotyledons. Sometimes there seems to be an indication
of an approximate tristichous arrangement of the leaves,
but it is not very close, and is probably accidental. The
phyllotaxy is probably a complex many-ranked one (very
likely the same as in Tristicha), and sufficient leaves are not
formed to enable one to recognize it. In a very short time
it is evident that the growth of the seedling is taking place
in a lateral direction, that a marked dorsiventrality is appear-
ing, and that there is no more development of leaves in the
vertical direction at all. The direction of growth is nearly
always at right angles to the plane of the cotyledons, and
horizontal along the rock. All the newly-formed leaves are
now seen to lean in one direction, that in which the apex is
growing. PI. IX., fig. 7, shows a seedling with seven leaves
besides the cotyledons, in which this growth is already
becoming indicated, and in fig. 9 it is clearly evident. In
this specimen, which is seen a little obliquely from above,
one of the cotyledons is absent ; there are four larger lateral
leaves somewhat divergent, and a considerable number of
small linear leaves all pointing nearly in the same forward
direction, while the tuberous hypocotyl is also evidently
elongating. The same stage is shown in side view in fig. 8,
and shows this growth of the hypocotyl into the “ shoot ”
thallus.
312 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
Once formed, the growing point extends itself along the
rock pretty rapidly. I have never been able to find that the
hypocotyl gives rise to more than one growing point, but in
the absence of large quantities of material it would be rash
to dogmatize. It will be well to point out at this stage one
of the many traps into which one is liable to fall in working
at these plants. Frequently I have obtained what I thought
to be seedlings, but which on thorough investigation on the
spot proved to be rejuvenescences of old exposed portions
of thallus, which, as we shall see, are very common. The
new growing points formed from these old thalli are exactly
like seedlings, and still more so when, as soon happens, the
old thallus from which they sprang shrivels away and is
washed off by the current.
The next stage that I have found is one about half way to
that figured in PI. IX., fig. 12, in which owing to branching the
seedling now has seven growing points. It will be observed
that the dimorphism of the leaves, which was beginning to
show itself even in the minute seedling of fig. 9, is now very
distinct ; there are large ovate leaves on the edges and
smaller linear ones on the upper surface of the growing tip,
the former in a single row on each side, the latter in a dense
irregular crowd. This is well shown in PI. IX., fig. 10, which
represents a particularly distinct case sometimes found.
The marginal leaves are parallel to each other, each with its
lower and anterior edge overlapped by the upper and pos-
terior edge of the next younger leaf. At the very tip the
lateral leaves are more or less upcurved, to some extent
covering the apex.
PI. XI., fig. 5, which is slightly diagrammatic, shows a
longitudinal section through a growing point in October.
The ‘‘apex,” or perhaps rather growing margin, of the
thallus is asymmetrical, the youngest cells being a little short
of, and above, the extreme tip, as shown at m. On the
under side of this tip there is a layer of collenchymatous
cells (coZ.), which (in the flabelliform thalli, at any rate)
split off as the growing point expands, and thus appear to
OF CEYLON AND INDIA.
313
behave like a root-cap, especially like the cap we shall
presently describe in Podostemon, Hydrobryum, &c. The
leaves can be seen forming, in large numbers on the upper
side, and in small on the lower.
There is no very definite arrangement, so far as I have
observed, in the meristem, but the peculiar structure of this
growing point requires further detailed investigation. The
meristem is small-celled, the cells near the growing point
being approximately isodiametric, while further back they
are elongated in the direction of growth of the thallus.
A desmogen strand of very narrow cells makes its appearance
in the middle of the thallus at a very little distance from
the tip. The dermatogen layer is fairly well indicated nearly
to the very apex. Rhizoid formation begins a little way
back from the tip, at the region where growth and movement
are probably completed or nearly so, by the outgrowth of the
epidermal cells in the usual way. Most of the rhizoids are
sloped a little forward in many specimens, as if there had
been a slight growth after they had become affixed to the
rock. Looking at a growing apex in surface view the
square shape of the cells near the apex is easily seen,
and the elongated shape of those further back is very prettily
indicated by the shape of their contained silica bodies.
These bodies are abundant in the surface layers of the
thallus, and have already been described in Tristicha (and
cf. PL XI., figs. 1, 3). Their presence makes section cutting
very troublesome.
To return to the large seedling mentioned (PL IX., fig. 1 2).
Its growing points are all of the pattern described above, and
have pretty evidently all sprung from the single original one
by branching. The branching of the thalli is so near the tip,
and the two branches are so nearly alike, that it looks like a
dichotomy ; but though I have not been able to definitely
decide this point, I think it is really a case of lateral branching.
The growing point itself is the centre of a cushion of tissue
thinning off towards the edges. When a branching takes
place, the two new growing points soon have a sort of valley
314 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
formed between them, the tissue there being thin, like that
at the edges of the original growing thallus. This thin tissue
soon ceases to grow in thickness or length, and presently
breaks, leaving an irregular gap between the two growing
points. The seedling figured shows various phases of this
process. It will be evident that the breaking away is
not a very definite process, being sometimes very marked,
sometimes hardly at all, as is indicated by the persistence of
the large marginal leaves along the greater part of the gap.
The cells of this thin tissue do not elongate so much as those
in the main course of the thallus, though they develop the
silica bodies.
In PI. IX., fig. 11, a specimen is shown in which the
branching is clearly visible at the tip, and it can be seen that
there is one of the large marginal leaves exactly central
between the branches, overlapped on each side by their
marginal leaves.
From this stage onwards the growth is evidently of the
same type, and in August and September, by which time it
is again possible to get material, the plants have reached a
considerable size ; several such plants are shown in the
photograph reproduced in PI. XII. The irregular but still
definite manner in which the thallus has grown to this size
is clearly evident. Several plants, it will be noticed, are
growing side by side, and are already commencing to inter-
fere with one another and to overlap, so that in the dry
weather a few months later it will be very difficult to
disentangle their mode of growth. It will also be noticed
that as the plants get larger their apical portions get broader
and more fan-like in shape. This is the expression of the
fact that the branching or formation of new apices becomes
more and more rapid in proportion to the rate at which by
the elongation of the thallus these apices become separated
each on its own piece of thallus. Along the edge of each of
these fans there is a very large number of growing points,
as shown in PI. XI., fig. 2, which represents a portion of the
edge of the thallus of the common Ceylon form at the season
OF CEYLON AND INDIA.
315
when the flowers develop. This figure also shows the course
of the vascular bundles in the thallus. It will be observed
that there is one to each growing point, and that they unite
as they pass backwards. The small linear upper sets of
leaves of each growing point run backwards for some distance
along the course of the bundles, but at a distance of about
10 mm. from the tip they have usually mostly fallen ofif,
though I have occasionally found specimens with leaves still
visible at 50-75 mm. from the tip. Their somewhat fugacious
nature, and their small size, is probably the reason why they
have never been noticed by any one but Tulasne. They do
not go far enough back to be among the older flowering parts
which are those usually gathered by collectors.
The flabelliform type of margin or apex, which has
just been described for the common Ceylon form (L. z.
Gardneriana), is not the only one to be found. The other
Ceylon variety (Parkiniana) has ribbon-like apices, with one
or few growing points upon them, as figured in PI. IX., figs.
13, 14. Among the Indian material both types also occur.
Bearing in mind the form of the apices of the seedling, we
shall perhaps be justified in regarding this second type as
the more primitive ; the figure in PI. XII. perhaps represents
not only the ontogeny of the individual flabelliform. tip, but
also its phylogeny. The tip of a growing point in material
from Khandala has been figured by Goebel (13, p. 169) and
Warming (42, IV., p. 160), and I have verified their obser-
vations upon material collected by myself. The general
size of the growing points is somewhat larger, and especially
the leaves, which, instead of the 2-6 mm. of the Ceylon
forms, may have a length of 8-15 mm., but the arrangement
of the parts is similar. Very commonly in these northern
forms the growing point is turned up clear of the rock, at
least at the time that it begins to form the flowers, for the
young stages have not been seen. This phenomenon may be
simply due to the crowding of the apices together ; very
often two large lobes of thallus are separated by a kind of
hedge of leaves due to the erectness of the growing points
(43)
316 WILLIS : MORPHOLOGY OP THE PODOSTEMACEÆ
along the two closely approximated margins. This is faintly
indicated even in the Ceylon plant figured in PL XII., where
in the second plant from the top the “ hedges ” can be seen
as white lines through the large fan-like lobe.
The leaves require but little notice. They have been
figured and described by Warming (42, IV., pp. 160, 163).
They vary much in size in different forms, as well as on the
upper surface and sides of the same growing point, but they
are always moss-like leaves of very delicate texture, like the
leaves on the ramuli of Tristicha ramosissima. They vary
according to position on the thallus from linear to ovate or
almost triangular, and in the lower part of each is a white
mesial line showing very clearly in the living leaf ; the line
has the corresponding linear or triangular shape to the leaf
in which it occurs. Microscopic observation shows that the
line represents a strand of narrow cells running up the
middle of the leaf, and probably having some vascular
function (PI. XI., fig. 3), while cross section of the leaf shows
that in this central part the leaf is several cells thick, though
the margin is only one cell thick (XI., 4). Silica bodies
are common in the angles of the leaf -cells (PI. XI., fig. 3).
The thallus itself has a simple structure as seen in cross
section (PL XL, fig. 1). On the upper surface is a well-
marked more or less columnar epidermis, and under this,
forming the body of the thallus, is a large-celled parenchyma.
Many of its cells contain large silica bodies, more especially
those of the two uppermost layers. On the lower surface is
a less conspicuous epidermis, bearing the rhizoids. The cells
contain chlorophyll, and large quantities of starch. A little
below the middle the vascular bundles take their course,
which has already been described as seen in surface view (PL
XI., fig. 2). The structure of the bundle is simple ; it consists
of a few small cells, of which the upper are apparently of
phlcem nature, the lower xy 1 em. I have not yet investigated
the development of the bundle in detail.
We have next to consider the development of the peculiar
rosettes of leaves, which are found upon the older thalli, and
OF CEYLON AND INDIA.
317
which Warming showed to be endogenously formed. These
soon begin to make their appearance. The first but not the
only place in which they appear is commonly in the angles
between the branches of the thalli, as shown in PL XI., fig 2.
They appear in this position without any actual separation
or splitting of the thallus taking place. They also very
commonly spring from the thin marginal portions of the
thallus or its branches. They appear at first as little swell-
ings of the upper surface of the thallus, formed, as Warming
has described, by an endogenous growth under about two
layers of cells, and soon burst out through the upper epider-
mis. The diagrammatic sections in PL XI., figs. 6, 7, show
how the growing point is formed by transverse divisions of
the parenchyma cells between two vascular bundles ; the
appearance of the tissue under the rosette is not unlike that
under the cups of an Æcidium. The growing apex itself
does not come above the surface of the thallus, but the leaves
burst through and spread themselves out in a symmetrical
rosette. They are usually linear or lanceolate in form, and
have the same structure as the leaves at the growing point
of the thallus itself. Beyond what has been described, no
order seems to be followed in the development of these
secondary shoots upon the thallus, new ones often appearing
ajnong the old. On the whole, however, they appear in
longitudinal rows down the centre of the strips of thallus ;
in the narrow forms only one or two rows may be found
(XI., 9), in the more flabelliform several (XI,, 2).
When these endogenous shoots appear upon the upper sur-
face of the thallus, but not actually at the edge of it, they
show radial symmetry, as just described, but when, as in
rejuvenescence especially is not infrequently the case, they
form upon the margin of the thallus, they show a distinct
dorsiventrality from the very first. PL XI., fig. 8, shows such
a shoot soon after its appearance. It rapidly develops into
a growing point exactly like that at the tip of the main
thallus, and may commence at once to form new branches of
the thallus, or to form a flower, according to the time of year.
318 WILLIS : MORPHOLOaY OF THE PODOSTEMACEÆ
We have now brought the life-history down to the time
when the dry September weather allows of the study of the
morphology of the young plants after the high water level
of the earlier part of the monsoon. At this season they have
usually reached a considerable size, like the specimens figured
in PL XII., often 10-20 cm. in diameter, with well-grown fan
apices and with numerous rosettes of leaves upon the older
parts; only on one occasion have I seen any sign of flowers,
and then I found a few ready to open ; but as a rule the
development of the flowers begins much later in Ceylon, for
I have found no sign of it in material gathered so late as the
first week in November. Owing to the shorter wet season
in India it is evident that the development must begin
earlier there. Now, unfortunately, the high water of the
other monsoon sets in, and it is impossible to get material,
and when the water-level once more sinks the development
of the flowers is already finished, so that this stage of the
life-history must at present be left to supposition. It is not
difficult, however, by a comparison of the growing tips at
the two periods, to infer with some probability in what
manner the development of the cupule, at any rate, takes
place.
The flowers are almost always marginal or apical, formed
from the characteristic apical growing points with their
dimorphic leaves ; only very rarely indeed, is a flower found
on the pper surface of the thallus, when it may possibly
have been formed from the growing point of a leaf rosette.
Very commonly, as indicated in PI. XL, fig. 2, each growing
point in the large fan forms a flower, and almost all the more
lateral apices form flowers in nearly all cases. Each apex
forms one flower only, which stands more or less erect on a
pedicel emerging from a horizontal or ascending cupule,
which is covered with slender bristle-like leaves closely
packed together on the upper side, has a membranous leafless
nearly flat under surface, and a few larger leaves on thetwm
lower margins, showing in fact the same general dimorphism
and arrangement of the leaves as the original vegetative apex.
OF CEYLON AND INDIA.
319
The upper side of the cupule is thicker thaiitLe lower. The
pedicel starts from the base of the internal leafless hollow.
There is much variation among the different forms in
regard to the form and depth of the^cupule and the length of
the leaves on it. PL XIII., fig. 5, shows the tj^pical cupule of
the common Ceylon form, just before it is exposed to the air
by the fall of the water-level. In this form, and also in the
other Ceylon form (Parkiniana, PI. XL, fig. 9), the cupule,
as a rule, remains closed over the tip of the flower until
exposure to air, but the opening at the apex is not completely
impervious. When cut in transverse section this cupule
shows the general appearance of the diagrammatic sketch in
PL XIII., fig 9, This shows the thick upper and thin under-
sides, the two lateral vascular bundles, the leaves on the upper
wall and margins, and the general orientation of the enclosed
flower. PL XIII., fig. 8, shows the same in longitudinal
section, and shows that the pedicel in , this form does not
elongate till the emergence of the flower.
Examining the forms from the Bombay districts (L. z.
konkanica) we find the cupule in what may perhaps be re-
garded as a more primitive condition ; it is thinner, less
obtrusively axial in nature, and opens sooner, while the
flowers are still submerged. Such a cupule with a nearly
ripe fruit is shown in PL XIII., fig. I, and in section in fig. 2.
What appears to be the case, so far as I have been able to
observe, in these plants is this. The cupule forms simul-
taneously with the flowei-, but not so rapidly as in the Ceylon
forms, so that it never really shuts in the flower at the apex,
and even after the flower has emerged to some extent the
cupule continues to grow. Hence it comes that those who
have examined only young buds have often supposed this
plant to have no cupule ; in this stage the cupule is often all
but evanescent, and sometimes the flower may ripen its fruit
with the cupule in this condition. Other growing points on
the same plant, again, may have very long cupules indeed.
Non-observance of these differences (which were correctly
320 WILLIS ; MORPHOLOGY OF THE PODOSTBMACEÆ
noticed by Tulasne) has led to several errors in the descrip-
tions of these plants. Thus Wight divided Law’s material
into two species (pedunculosa and foliosa) on the differences
of the apices of one plant, and Warming has described as
L. foliosa material collected by Goebel at Khandala, in
which I have found the majority of the apices to have well-
developed cupules. Hence, as I have already mentioned
in a preceding paper, the species L. foliosa (Wight), Wmg.,
is quite untenable.
The more southern Indian forms (L. z. malabarica), collect-
ed by Mr. Barber, show very similar cupules, on the whole
deeper and more like those of the Ceylon forms (PI. XIII.,
fig. 3).
Looking back to the structure of the vegetative growing
point already described, it is easy to see how these cupules
may have arisen by an invagination of the apical point
during the formation of the flower. The flower, in Ceylon,
only emerges from the cupule after the fall of the water has
exposed it to the air, but in India I commonly found the
flowers growing upwards through shallow water, and in one
case at least I found flowers which being submerged at some
depth in a pot-hole could not grow stalks long enough to
bring them into the air, and had become fertilized under
water, though not perhaps strictly cleistogamously, as the
perianth segments had separated. Many ripe fruits had been
formed in the pot-hole. True cleistogamic flowers occur in
Podostemon Barberi, as will be described below.
The flower emerges — in the Indian forms often with a
nodding stalk — on a pedicel, which at anthesis is about 2-5
mm. long, usually erect, whatever may be the slope of the
rock, and which has a clear pellucid cortex through which
the central vascular bundle is visible (PI. XIII., fig. 3). At
the base the pedicel is continuous with the thallus tissue.
Although the thallus and the cupule are so very dorsi-
ventral, and though the dorsiventrality of the thallus appears
at so early an age, yet the flower itself shows radial symmetry,
in spite of its having developed in a more or less horizontal
OF CEYLON AND INDIA.
321
position. Its structure is simple, and has been sufficiently
described by previous authors. It has a sepaloid perianth,
which is divided part of its height into three segments.
This at anthesis stands up closely round the essential organs
and holds the three slightly exserted stamens with their
introrse anthers very close to the stigmas, so that autogamy
is the rule (PI. XIII., fig. 3). The flower is very small and
inconspicuous, and the loose powdery pollen is easily blown
about by wind, and as the stigmas are receptive a little time
before the dehiscence of the anthers, and the flowers are
very close together on the rocks, there is a slight chance of
an occasional cross, but in general the fertilization is from the
pollen of the same flower. Every flower, as a rule, sets a full
complement of seed. The ovary is trilocular (PI. XIII., 10),
with a thick central placenta and very many minute ovules,
and crowned by three short papillose stigmas. The ovule
shows a simple structure like that described by Warming
for Tristicha hypnoides.
The ripening of the seed is very rapid, and in about a
week or ten days after the opening of the flower the
capsules have generally shed their seeds. The ovary wall
in the flower is smooth, and the ovary triangular, with
three bundles visible in the wall of each carpel. After
fertilization these thicken and lignify into stout ribs, as
Warming has described. No dehiscence ribs form at the
junction of the carpelsin this form, as is the case in the
Dicræas, &c., to be described below. The capsule (PI.
XIII., fig. 7) has nine well-marked ribs, with the thin wall
depressed between them. At the same time the tissue
round the vascular bundle in the centre of the pedicel (PI.
XIII., fig. 4) becomes strongly lignified as in Tristicha,
while the pellucid outer cortex falls away, leaving the fruit
on a rather longer pedicel than the flower, but on a slender
and very elastic one, as compared with the stout non-elastic
stalk of the flower. The lignification of the outer part of
the vascular bundle usually proceeds some little distance
back into the tissue of the thallus itself.
322 WILLIS ; MORPHOLOGY OP THE PODOSTEMACEÆ
The length of the pedicel of the fruit is very variable,
especially in the more northern forms ; I have found
pedicels from 6 to 25 mm. long on one plant. What deter-
mines the length I cannot say, but incline to think that it
is partly the rate of fall of the water-level, this again partly
depending on slope of rock, &c. In determining varietal
distinctions, the average length of a large number of fruit
stalks should be taken. A half- ripe fruit also may ripen
after gathering without lengthening its stalk.
Rejuvenescence, — One of the most important features in
the general life-history of this, as of other species of the
order, is its almost infinite capacity for rejuvenescence of
its thallus from any part thereof which remains alive. So
long as any portion, however small, of the thallus is alive, it
seems capable, if submerged in favourable conditions, of
forming anew growing point and re-commencing its growth.
The new apex forms near the edge of the old or damaged
portion of thallus, and begins at once to grow outwards
horizontally, showing the same structure and development
as the growing points previously described. An example
of such a new growing point is shown in PI. XI., fig. 8. The
importance of this capacity to the plant is very great. The
ordinary growing points soon die if exposed to air by the
fall of the water, and are also liable to damage from larvæ
of insects and in other ways. The damage to the plant is,
however, only temporary, as new apices are at once formed
and the growth goes on again. It is largely owing to the
extensive formation of these secondary growing points that
the morphology of the mature thallus is so complex and
difficult to unravel. The secondary growing apices may form
flowers, just like the primary ones.
Owing to this capacity for rejuvenescence, the plant can
hardly be called strictly annual. As the water falls in the
dry season, the bulk of the plant, or all of it in some cases
if the fall be very rapid or the rock horizontal, becomes
exposed to the air, and if it so remain for more than perhaps
a week or ten days, it often or usually dies, having in the
OF CEYLON AND INDIA.
323
meanwhile opened its flowers and ripened its seeds. Any
growing point, however, that lies near the lower side of the
plant will now tend to grow downwards along the rock, and
may succeed in growing downwards faster than the water
falls, so that the plant may thus survive the dry season with
its original growing points. Oftener, perhaps, this survival
is by means of new apices, formed from exposed thallus
at a time when a shower has caused a temporary rise of the
water and so covered again part of the old thallus. Especially
may this be seen in April, when the water reaches its lowest
and then rises again vdth the rains of the little monsoon.
Great numbers of young plants may then be seen, which
are easily mistaken for seedlings. Closer examination,
however, shows them to be formed from the old thalli
rejuvenescing in the newly submerged portions."*^ Often
new growing points may be seen on portions of thallus
bearing ripe fruits. Later in the season these young plants
are still more like seedlings, as the dead portions of the
thallus are now decayed and removed, so that nothing
remains but the star-like plant formed by the branching
of the new growing point, and closely resembling a seedling.
Physiology. — A few notes on certain physiological features
exhibited by these plants must now be given. A general
discussion of the physiology of the group will be given later.
The chief necessary conditions of life for Lawia zeylanica
seem to be a rigid substratum on which to grow, preferably
smooth rock, though it will grow on logs of wood or twigs
fixed among the rocks ; (2) rapidly running water covering
the plant ; (3) sufficient light. Like most of the other mem-
bers of the order, it will not grow well, and in nature is very
rarely found, in shady places on the rivers ; usually it occurs
in spots which for part of the day at any rate are fully
illuminated by the sun.
The cultivation of this species is very difficult. I have
succeeded in keeping it alive for some weeks in running
E.g., on 28th April, 1898, I found such specimens 2-4 cm, long.
(44)
324 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
water on the verandah of the laboratory, but not in getting
it to continue its apical growth in such circumstances.
The sensitiveness of the thallus to contact with the
substratum is very marked, and the thallus follows every
irregularity of the rock on which it grows. As, on the
whole, the growing points tend more to grow upwards than to
grow in other directions, it may be perhaps concluded that
they have a certain sensitiveness to light or gravity. The
growing point itself shows a marked dorsi ventral structure,
but the flower which terminates it does not, though it is
developr^d in a more or less horizontal position. Corres-
ponding to its radial symmetry of structure, it no sooner
escapes from the cupule than it turns upwards, and assumes
a nearly erect position.
Though adhering very closely to the rock, the plant does
not seem dependent on its root-hairs for food, and it is doubt-
ful if it absorb much from the rock, unless it obtains from
it the great quantity of silica with which its cells are filled.
The use of this silica is questionable, but probably it has some
secondary value in preventing the too rapid drying up of
the thallus when exposed to the air. Assimilation probably
goes on in both leaves and thallus, and the relative areas
would indicate that there is more in the latter. At an early
period in the life-history, the thallus contains a very large
quantity of starch, and later in the year, when the flowering
shoots are developing, the amount stored up at their bases
and in their tissues is enormous. To this is to be ascribed
the very great rapidity with which the plant is able to open
its flowers and to ripen its seeds.
The deep red colour of the majority of the plants is very
striking, and when the water is low and the sun shining
brightly, it is not infrequent to see spots where Weddell’s
metaphor is not exaggerated, “ le fleuve semble, si l’on me
pardon l’expression, rouler sur un tapis de roses. ” The
universality of this colour in the plants of this order, as in
the deep-growing Florideæ, seems to point to a definite
physiological meaning of the colour.
OP CEYLON AND INDIA.
325
The exceedingly dwarf habit of the species, and its
wonderful capacity for rejuvenescence, are rather biological
than physiological features.
The classification of the Lawias is rendered exceedingly
difficult by the variability of many of the characters, such as
the length of leaves and pedicels, and the size or depth of
the cupules. Seven species have been recognized. Of these,
L. ramosissima is more allied to Tristicha, to which I have
transferred it. Historical evidence and comparison of forms
showed that L. pedunculosa was a synonym of L. longipes
and L. Lawii of L. pulchella. The evidence on which the
species L. foliosa was based has been demolished above.
Tulasne’s two Indian species are based on differences in
length of leaves and pedicels, but as I have found these
characters to vary enormously on the same plant, with depth
of water, age of plant, and other things, these two forms must
be united, and as further investigation shows, that their
differences from the Ceylon forms are probably not enough to
be specific, we are thus reduced to a single species, L.
zeylanica, Tul.
In Lawia we have evidently a highly modified plant to deal
with, and it is also evidently adapted to more peculiar con-
ditions than is Tristicha. At the same time it represents
distinctly a side line in the evolution of the Podostemaceæ,
and is the only plant with a shoot thallus in the highest group,
the Tristicheæ.
The thallus, or at any rate its upper side, is here pretty
evidently of “ shoot ” nature, and is without either roots or
haptera ; it is of the most extreme dorsiventrality, and lies
closely down upon the rock. It is thus well suited to con-
ditions which, as pointed out above, must have had an
important influence in the evolution of the Podostemaceæ ;
no conceivable force of current, apart from scour, can detach
it from the rock, and it can also, owing to its exceedingly
dwarf habit, survive so long as there is the least trickle of
326 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
water flowing over it. The actual distribution of this plant
seems to point to the fact that this character is of great
importance, and may be regarded perhaps as a direct adapta-
tion to life in shallow water, or water liable to become
shallow very rapidly. In Ceylon the species is distinctly
infrequent, and it becomes more and more common
proceeding further and further northwards through Western
India, i.e,^ in going into regions in which the season of deep
water becomes shorter and shorter. In the Khandala district
the plant is enormously abundant, and is found in little
brooks up to points quite near to their sources. In this district
the rainy season is very short, but there is a trickle of water
in the streams for some time, and the plants live in this.
This reasoning is supported by the further fact that the
other species of this district, Hydrobryum lichenoides, is
also extremely dwarf. We shall return to the matter
below.
The great capacity of rejuvenescence possessed by this plant
must also be of great advantage to it. It has an obvious
advantage over Tristicha in the matter of the risks caused by
fall of water-level at a non-lloriferous period ; exposure only
kills a small part of it, instead of perhaps several large and
complicated shoots, and owing to the fact that the thallus can
bud out again from any part, growth is at once resumed,
without the plant having suffered any serious loss. The
thallus itself, too, is more or less amphibious, and
can form new growing points even after considerable
exposure.
The endogenous secondary shoots are an interesting
feature, but are of course not to be regarded as homologous
with those of the root thalli of preceding genera.
The mature characters of the plant appear very early
indeed in the life-history, the dorsiventrality appearing
almost immediately after germination. On the other
hand, though the plant is so dorsiventral in its mor-
phology and anatomy, the flower shows no sign of this
character.
OF CEYLON AND INDIA.
327
P0D08TEM0N.
[Michx. ; Willis, Rev. Podost. Ind., Ann. Perad., I., p. 228.]
As already explained in the preceding paper, the Indian
flora contains no representatives of the systematic groups
which intervene between the Tristicheæ and the Eupodoste-
meæ, with the latter of which we have now to deal for the
rest of this paper. In the systematic order there followed
Podostemon comes after Dicræa, but for morphological
purposes it is perhaps simpler to take the former first, as in
it the secondary shoots are still of considerable size, and take
the chief part in assimilation, whereas in Dicræa and nearly
all the following genera they are very much reduced as
compared with the thallus.
As elsewhere explained, I have excluded from the genus
most of the species placed in it by Bentham, who is followed
by Hooker in his Flora of British India ; these species are
considered under Dicræa, Hydrobryum, Griffithella, and
Willisia. There can be little doubt that Bentham’s fusion of
all these went either too far or not far enough, for if these
be united with Podostemon there is no valid reason for
keeping either Mniopsis, Oserya, Devillea, or Ceratolacis
separate from it, whereas in the Indian species which he did
unite with Podostemon there is very great morphological
variety, as we shall see.
The genus, as now defined, is confined to America and
India, so far as our present knowledge goes, and includes in
India and Ceylon two very isolated forms, P. subulatus,
Gardn.,and P. Barberi, Willis, to which perhaps should also
be added Polypleurum Schmidtianum, Warming, lately
discovered in Siam.
Our knowledge of the morphology of this genus is com-
paratively good and complete, thanks to the labours of Prof.
Warming, who has described the thallus, secondary shoots,
and flowers, in many different species from America, and in
P. subulatus from Ceylon material collected by Trimen. The
American species mostly show a creeping thread-like thallus
328 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
of root nature, very similar to that of Tristicha ramosissima,
attached to the rock by hairs and haptera, and with endo-
genous branching and a root-cap. From this are given off,
often more or less paired, the secondary shoots, which grow
upwards to a greater or less height, bear distichous leaves,
branch more or less often, and bear several or many flowers.
The early stages in the life-history, even of such com-
paratively accessible forms as the Ohio P. Ceratophyllum, are
however completely unknown.
PodostemoBi subulstftus, Gardn.
(Plates XIV.-XVI.)
I have investigated this species chiefly at Hakinda, where
it is abundant, and I have also found a variety of it in the
Anamalai mountains. The latter is characterized by smaller
habit and longer pedicels, and may perhaps ultimately prove
specifically different.
Habitat. — In Ceylon the plant occurs in the Mahaweli-
ganga and its tributaries, between 1,500 and 2,000 feet. At
Hakinda it is common, and presents a very striking appear-
ance, especially when seen in the height of the vegetative
season in September. It grows like the rest only on a rock
substratum, and is rarely found in rapidly moving or broken
water, but occurs in great masses in the more quiet bays in
the rocks among the rapids, where the eddy keeps the water
in constant slow movement past the plants, waving the long-
red or green leaves to and fro, like Algæ on the rocks of a
quiet seabeach. Like the rest it is never found on unstable
substrata or in stagnant water. It is very often mixed with
Dicræa elongata, more rarely with D. stylosa, vars. laciniata
or fucoides, or with Farmeria metzgerioides, and occasionally,
at places where an eddy returns into the main rush of the
stream, it occurs mixed with Lawia zeylanica or Hydrobryum
olivaceum. Its habitat in the one place where I found it in
the Anamalais was very similar ; it was not mixed with any
other species, but close to it, at the edge of the eddy in which
it grew, were Willisia selaginoides and a Hydrobryum.
OF CEYLON AND INDIA.
329
Dry Season Appearance. — The left-hand figiirr in PI.
XIV. shows a stone covered with this species in the condition
in which it is found when the water has fallen away from
it and the fruit is ripe. The rock is covered with a little
forest of small fruit-hearing stems, 1-3 cm. high, marked
with the scars of fallen leaves (PI. XVI., fig. 11). The fruits
are shortly stalked, and when open have lost one valve.
Often the thallus may still be seen as a thin thread on the
rocks between the stems. An examination of material
taken at this time from below the water (such as was
studied by Prof. Warming) shows that the shoots bear long
subulate leaves, in the lower axils of some of which the
floral shoots arise. These leaves shrivel and drop off at
once on exposure to the air. The structure of the shoots at
this period has been very fully described by Warming, and
the following account will deal mainly with the earlier
parts of the life-history.
Germination and Life History. — The seed is like that of
the species already described, but has a slightly different
embryo, the cotyledons being short and subulate, instead
of thin and crumpled. There is a mucilaginous layer in
the seed coat, as usual. When wetted the seed swells ; the
hypocotyl emerges and bends downwards, and becomes at
once attached to the substratum by rhizoids, as in Lawia ; it
then becomes slightly tuberous, and the cotyledons expand.
They are opposite to one another, and grow to a length of
about 4 mm. The upper surface is slightly channelled at
the lower part, but is not distinctly sheathing like the
mature leaves. The cotyledons are not hairy, though the
ordinary leaves are so.
The primary axis develops only to a very slight degree ;
it remains more or less erect, but the cotyledons and leaves
usually bend over towards one side. The leaves soon
appear (PL XVI., fig. 1), the first two being markedly
alternate, and approximately in the same plane as the
cotyledons. The second leaf, as seen in the figures,
develops within the sheath of the first. About six is the
330 WILLIS : MORPHOLOGY OF THE PODOSTBMAOBÆ
greatest number of leaves that I have observe rl to develop
upon any primary axis that I have seen, and the leaves are
comparatively small. All the leaves appear, in one plane,
i.e.y are distichous, with the cotyledons forming the lowest
member of each rank. The leaves are hairy on the upper
surfaces, and are usually, especially when young, of a deep
reddish colour from the presence of anthocyan in the
epidermal cells.
The thallus appears almost at once (PI. XVI., fig. 2)
growing out along the surface of the rock as a little papilla
emerging from the basal portion of the hypocotyl, usually
at right angles to the plane of the cotyledons. Its develop-
ment is endogenous. Once formed, its growth is rapid ;
it becomes attached to the rock by rhizoids and occasionally
by haptera. Its growing point shows a peculiar structure,
details of which must be reserved for future description.
It has no deciduous root-cap whatever, but the extreme tip
is formed by a collenchymatous layer beyond the initial
meristem, similar to, but less marked than, that which we
shall subsequently find in Hydrobryum and Farmeria.
The “ epidermal ” tissue (if such it be) of the maturer
parts of the thallus can be distinctly traced in surface view
round the actual apex, being there continued by the collen-
chyma. The extreme apical zone shows cells dividing in
all directions ; then follows a zone of mainly transverse
division, and then one of elongation.
Very soon the thallus begins to branch. This it does
very close indeed to the apex, and as might perhaps be
expected from the construction of the latter, it branches
exogenously. In PI. XVI., figs. 4, 5, the tip of a thallus is
shown which has just branched, and other figures on the
same plate also illustrate the branching. The branch forms
close to the apex, so closely indeed at times that until it is
observed that the endogenous shoot is also forming
between it and the apex, it looks almost as if formed by an
actual dichotomy ; in reality, however, the branching appears
to be lateral. The shoot seems to be considerably developed
OP CEYLON AND INDIA.
331
before the branching begins. One arm of the two thus
formed usually grows much more rapidly than the other.
At the next branching this is repeated on the other side,
and so on. The shorter branch of the two, as a rule, does
not grow out very far, and often remains very short, as in
the portion of a young plant figured in PL XYI., fig. 6,
where two of the three branches shown have simply formed
short projections, fastened to the rock in the usual way.
The similarity of this figure to that of the alga Bostrychia
Moritziana as figured by Goebel (Flora, 83, 1897, p. 436) is
of the most striking description. The alga, it is worth
noting here, and the subject will be considered again below,
occurs in the same place with Oenone Imthurni, one of the
Podostemaceæ of British Guiana. The frequently recurring
similarity of the Podostemaceæ with the Algæ of moving
water can scarcely be altogether “accidental.”
When one of the lateral branches thus formed branches
again, the first secondary branch is always upon the
basiscopic side of the primary. The thallus becomes more
or less zigzag in its construction, the two branches diverging
at every node. In transverse section (PL XVI., fig. 7) it
shows an almost cylindrical shape, fiattened on the lower
side, upon which is a kind of continuous foot, formed by
the mass of root-hairs or rhizoids. In the central portion,
but a little towards the lower side, is a vascular bundle of
very simple construction, described by Warming, who
states that there are sometimes two xylem groups towards
the lower side, characterized by spiral vessels.
Leafy secondary shoots are also developed upon the
thallus, and soon form, as in Tristicha, the most important
part of the plant. One shoot forms at each branching of the
thallus. PL XYI., fig. 3, shows the first leafy shoot appearing
very early in the life of the thallus, and fig. 4 shows it more
clearly, with its endogenous origin. The first leaf that
breaks through is usually towards the basiscopic side, as
there shown, and the other leaves follow in distichous order.
The first leaves are small, but the shoot grows rapidly, and
(45)
332 WILLIS : MORPHOLOGY OF THE PODOSTEMACBÆ
very soon leaves of 3-4 cm. long are produced, and after a
few months, in August or September, leaves may be
commonly found of as much as 30 cm. long. The shoot in
the Ceylon form usually grows more or less erectly, while
in the Anamalai form it is more prostrate, but otherwise the
two are very similar. Each shoot exhibits a certain amount
of dorsiventrality, the side next to the fork of the thallus
being the upper and slightly concave, while the other side
is lower and convex ; the basal portion of the shoot flattens
out into a sort of foot, figured by Warming, who has
described these shoots in detail. The dorsiventrality shows
itself in other ways also. PI. XYI., fig. 8, shows a section of
the top of a shoot through the leaves, and shows the slight
concavity of the upper side ; the vascular bundles form a
curved line, concave upwards, and the leaves themselves
show a distinct dorsiventrality, which resembles that of the
cupules of Lawia, in that the upper sides of the sheaths are
thicker than the lower. The lower sides, Avhile thinner, are
also larger in surface than the upper. We shall find these
characters much more pronounced in the leaves at the bases
of the flowers in Dicræa and other subsequent genera.
Occasionally in the Ceylon, and usually in the Anamalai
form, the stem branches at the base, but more commonly
there is no branching in the former until it has reached a
height of 0’5-2 cm. and is ready to flower. When fully
grown, one of these shoots may have a height of as much
as 3 cm. and bear as many as 20 leaves, each 20-30 cm. long.
The stems are produced near together upon the thallus,
so that when a plant is well-grown in September it may have
30 or 40 such stems, and the great mass of reddish-green
leaves waving in the water like seaweeds upon the rocks of
the coast, forms a very striking sight. A portion of such a
plant is figured in PI. XV. ; the separate shoots on the left
show their construction and the thallus joining them. The
leaves themselves have been described by Warming; it may
be noted that the very young ones are redder than the fully
groAvn leaves, that the upper surfaces are hairy, and that
OF CEYLON AND INDIA.
333
when taken from the water the leaves are very slimy, like
those of many other water plants. This particular species
is practically free from silica, and consequently shrivels
rapidly when taken out of the water, all but the stout erect
stems ; the leaves fall, and the thallus withers to a thread.
As the stems grow upwards the leaves fall off below,
leaving well-marked scars. Up to the beginning of the high
water of the north-east monsoon the stems (at Hakinda) are
mostly unbranched, and the leaves monothecous (i.e., with
sheath on the upper side only). Subsequently branching
occurs from the lower axils of dithecous leaves (with sheaths
on both upper and lower sides, cf. Warming), and a terminal
flower is developed on each branch. I have, as with the
other species, been unable to get material showing the
development of the flower. When the water falls in Decem-
ber the flowers may be seen fully developed in their spathes,
the latter not, however, opening until exposed by the fall of
the water. At this period the leaves are usually a good deal
smaller than in the earlier part of the vegetative season.
Except that the spathe is shorter and stouter, the general
appearance is very like that figured in PL XYII., fig. 1, for
Podostemon Barberi. The transverse section of the group of
leaves in PI. XVI., fig. 8, includes the section of the pedicel
of the flower, and an enlargement of this is shown in fig. 9.
It is nearly circular in section, has a small central vascular
bundle, a large cortex, and a columnar epidermis, especially
well marked on the “ upper ” side, so that even in the pedicel
there is a degree of dorsiventrality.
As soon as the water-level sinks sufficiently to expose the
plants, the leaves shrivel and drop away, and the spathes
open to allow the escape of the flowers. PI. XVI., fig. 10,
shows a flower fully open. It is without perianth, unless
the thread-like organs at either side of the common stalk of
the stamens be regarded as perianth ; it stands on a short
stout pedicel, nearly erect. On the lower side, Lö., the
convex side of the shoot, is the andrœceum, consisting of
two stamens on short partial filaments united upon a longer
334 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
common filament. The ovary is of two carpels, bilocular,
with a stout axile placenta, thin septa, numerous ovules, and
two stigmas. In cross section it is very like that of P. Bar-
beri (PI. XVII., fig. 6), but the outer margin of the deciduous
tissue is not undulate, so that the ovary is not winged as in
that species. In all other respects it is the same, and the
description given below will suffice for both.
The stamens stand away from the stigmas, so that self-
pollination is not absolutely inevitable, though it seems by
far more common than a cross. The flower is anemophilous,
and produces large quantities of loose powdery pollen.
Practically all the flowers appear to set a full complement
of seed. The seeds are very small and soon ripe.
The development of the fruit is of particular interest, and
as it is very similar in all the following genera, it will save
repetition to describe this one in full. As in most of these
plants, the seed is practically no larger than the ovule, and
the fruit than the ovary, as is shown in PI. XVI., figs. 10, 12,
and 13, which are all on the same scale. The pedicel, as
already described, has a thick cortex, which, as in Lawia, is
deciduous, leaving the now woody central tissue as a slender
but much more elastic stalk to the ripe fruit. With this
deciduous tissue there falls away also the outer tissue of the
ovary, as indicated for P. Barberi in PI. XVII., figs. 6 and 9,
leaving to the ripe fruit only the vascular tissues, which
become strongly lignified and form the ribs, and the inner-
most two layers of cells, already cutinized in the ovary,
which form the wall of the ripe fruit, and are very smooth
and shiny within, but rougher on the external surface where
the other cells have fallen away. The outer layer of the two
also becomes strongly lignified. The ripe fruit is thus, if
anything, rather smaller than the ovary, but with clearly
marked ribs.
The distinction between the genera Podostemon and
Dicræa is partly based on the mode of dehiscence. In the
former the two lobes are unequal, and one falls away leaving
the other on the stalk, while in the latter both are equal and
OF CEYLON AND INDIA.
335
persist on the stalk. The figure shows the construction of
the typical Podostemon fruit. As the cross section of that
of P. Barberi shows, there are eight ribs, of which the septal
ribs are double ; dehiscence occurs by their splitting and
the falling away of the septa. In Podostemon the ribs on
the one half of the fruit run down into the pedicel (XVI., 12),
while those of the other and deciduous half do not. The
halves of the dehiscence rib unite with the rib next to them
on each lobe. The process will be easily understood by a
reference to the figures, which may be compared with those
of Dicræa.
The capsules open as soon as ripe if the air be dry, and the
plant now presents the usual dry season appearance figured
in PI. XIV., fig. 1, in which the persistent lobes of the open
fruits can be clearly seen.
Rejuvenescence. — Like Lawia, this species exhibits a great
power of rejuvenescence, forming new shoots with great
rapidity if any necessity arise, as through injury. When,
as not infrequently happens in August or September, the
water falls rather low during the vegetative season, many of
the leafy shoots and parts of the thallus become exposed to
the air and die. As soon as the water rises again, however,
the tip of the thallus forms a new growing point, growth is
resumed, and is soon as free as ever. This process also
goes on in the dry season of February and March, when the
bulk of the plant is killed by exposure ; branches of the
thallus are formed and grow downwards over the rocks
more rapidly than the water falls, so that the whole of the
plant is not killed. It is rather hard, consequently, to say
whether this plant, or indeed any of these plants, should be
strictly termed annual. The new shoots formed in the dry
weather of the early part of the year do not flower till the
following dry season.
This species is fairly well represented in herbaria : most
of the specimens show leaves in situ on the stems, but only
rarely show the thallus.
336 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
Ecology, — In general this is much the same as in Lawia.
The plant is larger and less firmly fastened to the rock, and
affects less rapid water. Its habit is less dwarf, and among
the Indian Eupodostemeæ it is perhaps the one which
approaches nearest to Tristicha in general habit of thallus
and shoot, but it shows an advance on that genus in the
greater dwarfing of the shoots, and makes up for great
number of leaves by larger size of the few that occur ; it is
thus able to live in shallow water with perhaps greater
safety than Tristicha ramosissima. It still shows, what is
rare in the succeeding genera, the assimilation chiefly per-
formed by the leaves and not by the thallus. Vast quantities
of starch are, as usual, accumulated in the thallus, the shoots,
and the pedicel and placentae of the flower.
Podostemon Barberi, Willis.
(Plate XVIL)
This very interesting new species, more closely allied to
the preceding than to any other known species of the genus,
but differing in its flat thallus and its single-staminate
cleistogamic flower, as well as in a few other points, was
discovered at Beltangadi in S. Kanara by my friend Mr.
C. A. Barber, after whom I have named it.
Habitat. — It occurs in a rapid, but not violent stream, and
is mixed with Grifflthella Hookeriana, var. Willisiana.
Dry Season A'ppearance. — The plant in the fruiting con-
dition has some similarity to P. subulatus, but the fruits
have much longer stalks, and the secondary shoots are very
short. The flowers are easily seen to be cleistogamic, the
spathes not opening until the fruits are nearly ripe and often
hanging in wisps about the pedicels.
Mature Structure. — The germination and early life-history
of this plant have yet to be studied ; the only existing
material is that collected by Mr. Barber at the end of the
season. Probably, however, the general features of the life-
history are much the same as in the preceding species.
OF CEYLON AND INDIA.
337
The thallus, however, is very different ; instead of being
nearly cylindrical, it is flat, narrowly elliptical in section,
abont 1 mm. thick in the middle, and tapering oft* to the
sides (PI. XVII., flgs. 1, 3). Its breadth is usually about
3 mm., but in one or two cases the thallus is broader, and
almost seems as if it tended to become crustaceous, like that
of some of the Hydrobryums. It is frequently branched,
the branching being all in the horizontal plane and exoge-
nous. Along the middle of the lower sidé runs what may
almost be called a continuous foot, composed of root-hairs
closely matted and mixed with a gummy excretion as already
described for Tristicha.
In the middle of the thallus, as seen in cross section (PI.
XVII., flg. 3), is a vascular bundle like that of the preceding-
species, while the wings or sides of the thallus, as seen in
fig. 4, are composed of large-celled parenchymatous tissue,
with epidermis above and below. I have not seen in this
species any sign of the tangential division of the cortex of
the thallus that is so marked a feature in the very similar
thallus of some of the Dicræas. Nor do the margins of the
thallus seem to fall away here as in some of the latter.
The secondary shoots arise on the upper margins of the
thallus, in the angles of the branching, just as in P. subula-
tus. They differ from the latter chiefly in the smaller size
of the leaves (^.e., so far as can be judged from specimens
taken at the end of the year ; during the purely vegetative
season they may quite well be much larger), and in th e fact
that the lower leaves do not fall away to expose a scar-
covered axis. The leaves in the specimens that I have seen
rarely exceed 5 cm. in length. In structure ond arrange-
ment on the stem they are like those of the preceding species,
and the branching takes place as there from the lower axils
of dithecous leaves.
Flowering occurs, as usual, at the beginning of the dry
season ; this species shows what one is surprised not to find
more commonly in the order, cleistogamic flowers, fertilized
338 WILLIS ; MORPHOLOGY OF THE PODOSTEMACEÆ
under water. What is the stimulus causing maturity in the
flower has yet to he determined, whether greater exposure to
light, more dissolved air in the water, reduced water-pressure,
or other cause. The spathe (PI. XVII., figs. 2, 5) is adnate
below to the pedicel, is long and narrow, and encloses the
fiower until after fertilization. There is only one stamen,
which stands closely up against the stigmas. Germination
of the pollen takes place within the anther, but I have not
been able to make out how the tubes pass to the stigmas.
The ovary soon expands into the fruit after fertilization,
and the spathe, which for a time lengthens with the pedicel,
ultimately splits and falls away.
The six main ribs of the ovary, but not the two dehiscence
ribs, are provided with stout wings of delicate translucent
tissue, a feature of specific distinction between this species
and P. subulatus. PI. XVII., fig. 6, shows a cross section of
the ovary, and fig. 9 of the resulting fruit. These show, as
has already been partly described under P. subulatus, the
way in which the ribs of the fruit are formed. In cross
section the ovary shows a stout placenta, two thin septa,
and six wings, each with a vascular bundle. There are also
vascular bundles in the place where the dehiscence ribs are
to form. The inner epidermis in section shows a bright
yellow colour ; it consists (fig. 8) of long narrow cells with
very stout outer cutinized walls. Under it is a layer of
smaller somewhat thick- walled cells, and beneath this again,
towards the outer side of the ovary, is large-celled
parenchyma tissue. As the fruit ripens, this parenchyma
falls away, and the vascular bundles and the inner epidermis
and layer of cells under it alone persist. The inner layer of
cells and the outer parts of the bundles become lignified,
while the inner epidermis forms a smooth shiny inner layer
to the fruit wall. By the fall of the parenchymatous wings,
the two dehiscence ribs become almost as prominent as the
other six, and the two sides of each bundle become woody,
but not the central line, in which dehiscence takes place.
As seen in fig. 7 the stout inner epidermis also ends on either
OF CEYLON AND INDIA.
339
side of the line of dehiscence. The septa of the fruit break
away, and the dehiscence ribs split down the middle. As
in the preceding species, one valve of the fruit falls away,
the other persisting with its ribs decurrent into the
pedicel.
The genus Podostemon, from a morpho-ecological stand-
point, evidently falls into line with Tristicha, and it will be
of interest to consider its peculiarities as compared with that
genus, in connection with the difference in the general
conditions of life. On the whole, so far as can be judged,
Podostemon lives in very similar localities to those inhabited
by Tristicha, but the water is perhaps a little swifter.
Exposure to light, &c., and the other conditions considered
in the introduction seem much the same for both. Morpho-
logically, the chief feature in which Podostemon differs from
Tristicha is the great reduction of the secondary (and
probably the primary) shoots. The thallus is very similar,
though the root-cap is almost absent (as perhaps in Tristicha
hypnoides), and it is exogenously branched. The secondary
shoots, however, are much dwarfed, making up for the
reduction of the axis by the development of very large and
flexible leaves ; the number of flowers borne on them is also
much smaller. Thus, on the whole, probably the plants are
better able to take the risks of temporary fall of water-level,
and do not have to pay so dearly for exposure, and can better
stand shallow water. They also probably offer less resistance
to the current, and so can stand swifter water. The exten-
sion of leaf surface, obtained in Tristicha by the delicate
moss-like leaves of the ramuli, is here increased by the
formation of hairs on the surface of the leaves.
Dorsiventrality is considerable, and shows very markedly
in the flower, but less so in the vegetative parts than in
Lawia, and there is hardly any trace of it in the seedling.
The great development of the secondary shoots is accom-
panied in this plant by the curious feature common among
(46)
340 WILLT8 : MORPHOLOGY OF THE PODOSTEMACEÆ
the Eupodostemeæ, that the primary axis has become
insignificant, and takes no part in the bearing of fiowers.
P. Barberi shows the first sign we have yet met with of that
flattening of the tballns which is so marked in some of the
subsequent genera.
DICRÆA,
[(Du Pet. Tk.) Tul. ; Willis, Eev. Pod. lud,, Ann. Perad., I., p. 216.]
I employ this genus, as already explained, in Tulasne’s
sense. It is very abundant in all the districts of India and
Ceylon where Podostemaceæ ha^e been found, except the
more northern on each side of the continent. The species
are variable and polymorphic, and much detailed study is
needed to finally determine the specific limits ; by using
very broad distinctions I have included all my material
under five species, one of which is not, I think, really
separable from the rest.
This genus is characterized by the possession of very
remarkable root-thalli, two forms of which have already been
described by Prof. Warming, viz., those of D. elongata and
of D. stylosa, var. fucoides (D. algæformis. Trim., Wrng.),
both from material collected in Ceylon by Trimen at the
beginning of the dry season. These accounts are very full
and accurate, but a few corrections must be made in them,
and they do not deal with any of the early stages of the life-
history.
Oicræa elongata^ Tul.
(Plates XIV., and XVIII.-XX.)
Our knowledge of this species is mainly due to Warming,
who has described material collected at Hakinda by Trimen
in the dry season. My account is therefore chiefly confined
to the general life-history and the earlier stages, and to
the correction of a few errors in previous descriptions. I
have investigated the species at Hakinda, where it is very
abundant.
HaMtat. — This species is apparently confined to Ceylon,
where it occurs from the lowest slopes of the hills up to
OF CEYLON AND INDIA.
341
2,000 feet, especially in the Mahaweli and Kelani rivers. It
grows, as a rule, in rapidly running broken water, but not in
such violent streams as does Lawia. Its size and its manner
of drifting out with the current cause it to afford a much
greater hold for the water to pull upon. It is most often
found growing by itself, like most of these plants, but is
commonly found mixed with Podostemon subulatus or
Dicræa stylosa, var. laciniata, less often with var. fucoides,
or with Hydrobryum olivaceum. It is usually one of the
first to emerge as the water-level falls.
Dry Season Appearance. — As found in January or February
upon dry rocks the plant consists of numerous more or less
upright or projecting woody stalks, reaching to a length of
about 10-15 cm., with distichous stalked fruits borne at
intervals of about 1 cm. on each side. Along the rock, and
firmly attached to it, are Avoody creeping thalli, from Avhicîi
the erect ones spring. The erect stalks are also thalli, while
the lateral stalks borne upon them are the secondary shoots,
endogenously formed from them and bearing each one fruit
(PI. XIX.). Usually some plants may be found which are
still submerged, and then it can be seen that the erect woody
thalli are only the lower ends of very long thalli, which drift
out in the water to a length of as much as 50 cm., bearing
little fascicles of leaves towards the outer end, and fioral
shoots on the lower parts (PI. XIX., and cf. figs, in Tulasne
and Warming).
Germination and Life History. — The seed has the usual
structure, with a mucilaginous outer coat, and the germina-
tion takes place as in those species already described. The
cotyledons spread out approximately in the same plane, but
sometimes with a tendency to approach one another on one
side and thus give the embryo an asymmetrical structure.
Almost at once the leaves of the primary axis appear. The
first and second are usually, but apparently not always,
nearly at right angles to the plane of the cotyledons (PI.
XYIII., fig. 1), and they are folloAved by a small number of
other leaves, which are arranged in no very exact order or
342 WILLIS : MORPHOLOGY OP THE PODOSTEMACEÆ
phyllotaxy that I have been able to discover. The primary
axis is quite insignificant, and never reaches any size.
The thallus appears almost at once. The base of the
seedling axis forms a sort of tuber, as usual, which is fastened
to the rock by root-hairs and by the formation at a very early
stage of one or more exogenous hapfcera. The thallus
appears to be endogenous, but in some cases that I have
examined I was not able to satisfy myself that it was not
exogenous. Great care has to be taken not to confuse thallus
and hapteron in young seedlings ; the latter is always exo-
genous and is usually larger-celled. PI. XVIII., fig. 1, shoAvs a
young seedling with the thallus beginning to form. In fig. 3
a rather later stage is shown on a larger scale to indicate the
endogeny of the thallus, and the exogenous hapteron with
its larger cells, which are now apparently more or less fully
grown. A still later stage is shown in fig. 2. Here the
primary axis has reached almost its maximum development,
consisting of a few leaves on a very reduced stem. As soon
as the thallus is well established the primary axis becomes
quite an insignificant factor in the life of the plant, and in
older plants cannot usually be found at all. The thallus
groAVS along the rock, attached by root-hairs and by haptera
for a distance of one or tAvo millimetres, then checks in its
growth and forms a branch on the upper side, as seen in
figs. 2, 4, This branch is the commencement of a floating
or rather drifting thallus, Avhich extends upAvards into the
Avater and does not, as a rule, become attached to the rock at
any other point than its basal end. Usually the drifting
thallus seems to start from the creeping one at a gentle
foi-Avard slope or angle, as in fig. 2, but sometimes, as in
fig. 4, it starts at right angles. These drifting thalli form
the chief part of the mature plant, and being, like the
creeping thallus, green, do a large part, if not the bulk, of
the work of assimilation. The thallus groAvs rapidly by an
apical growing point (PI. XVIII., fig. 5), already described
by Warming, with a small terminal root-cap. A single
Avascular strand runs up the thallus to near the tip, and
OF CEYLON AND INDIA.
343
from it branches run to the endogenons secondary shoots,
which are produced in acropetal succession upon the
thallus, usually alternately, but often sub-oppositely. The
endogenous shoots have been fully described by Warming.
In contradistinction to what Ave have seen in Tristicha and
Podostemon, they are very much reduced ; the axis does not,
so long as the shoot remains vegetative, come above the thallus
surface at all, but remains almost evanescent, and the only
part appearing above the thallus is the leaves, of which
there is a small fascicle, about 4-8 in number, simple,
subulate, hairy on the upper surface, and not often more than
5-10 mm, long. The secondary shoots appear very near to
the tip, and are usually about 5-10 mm. apart ; they are
distichously arranged along the tAvo sides of the cylindrical
thallus, ^.e., on the sides which were lateral as the thallus
started from the creeping part.
The thallus branches freely upon or near to the rock, but
only rarely do the upper parts of the drifting thalli branch;
when this occurs, the branches are like the drifting part,
and lie more or less parallel with it. The branching, as a
rule, is exogenous, except when branches are formed on old
parts of the thalli, when they are commonly endogenous.
When a branch is formed, it usually creeps along the rock for
a short distance, and then gives rise, just like the primary
thallus of the seedling, to a drifting thallus. Thus ulti-
mately the plant forms a large number of drifting thalli and
a confused tangle of creeping thalli upon the rock. The
branching is less regular in this species than in most of the
others with which we have to deal ; the branches appear
more or less anywhere, instead of being in strict acropetal
succession along the thallus. By the end of September the
plant has reached its full vegetative growth, and is even
more striking as seen in shallow water than Podostemon
subulatus. PL XIX. shows such a plant ; there are many long
drifting thalli of lengths up to 50 or even 60 cm., waving out
in the water as the swift current passes by, but firmly held
by the creeping thalli at the base. The colour is green or red.
344 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
To turn for a moment to the internal structure of the
thallus at this stage, we find several features of interest,
details of which must as usual be postponed for a later paper.
In transverse section the thallus PI. XVIII., fig. 6) is
approximately circular or elliptical, with a single vascular
strand towards the lower side, surrounded by a cortex whose
cells are collenchymatous towards the centre, and with walls
thinner towards the epidermis. The vascular strand is
elongated in the horizontal plane (fig. 8), and near the tip of
the thallus is only two or three cells thick. At the ends may
be seen two small-celled groups, which may perhaps re-
present primary xylem, and the rest of the tissue is paren-
chymatous or phloem-like, especially on the upper side, where
the cells are narrower, and seem to represent primary
phloem ; no sign of any characteristic xylem markings
or of lignification is present. Further down in the thallus
meristematic activity is evident in the vascular strand. The
rather larger lower cells of the strand are found to be
dividing, chiefly at first in the horizontal plane, but later
also in the vertical or irregularly. A commencement of this
process is shown in fig. 9, and two cells, with their resulting
products, are shown in fig, 10 at a stage when the process is
nearly complete, and when the bundle has reached the stage
figured by Warming from January material. Besides this
growth in thickness in the bundle, there is also a corres-
ponding growth in the cortex. As shown in fig. 7, it begins
in the outer layers and on the upper side, and is at first
tangential, but very soon most of the cortex begins to
divide both tangentially and radially. Even at this stage,
though the bundle has reached some size, there seems to
be no lignification, nor any characteristic xylem elements,
though these may be seen later in the lower parts of the
bundle.
The heavy rains of the north-east monsoon now begin,
and by December, when the plant is again in shallow water,
the development of the flowers is practically complete,
though there is no sign of it in October. As might be
OF CEYLON AND INDIA.
345
expected, the floral shoots are simply metamorphosed
vegetative shoots, and there is no dimorphism in the sense
ordinarily employed in reference to this plant. Only on
extremely rare occasions does the whole number of the
secondary shoots become floriferoiis ; as a rule, about one-
third or a half of them is the maximum so metamorphosed.
The non-floriferous shoots are always towards the tip of the
thallus, the floriferoiis at the lower end, and not infrequently
some of the creeping thallus becomes floriferous also.
During this period of submergence the thallus has continued
to grow in thickness, mainly by meristematic activity in the
cortex, and it may now be as much as 4 mm. in thickness
near the base, tapering away to a mere thread towards
the tip. The leaves of the shoots towards the base usually
drop as the thallus grows longer, and presently the floral
shoots are formed at these points. What determines the
number of shoots that are to become floriferous, unless it be
the available amount of nutrition, is hard to say ; the boundary
between the flowering and non-flowering parts of the thallus
is very sharply marked. In the former the inner collenchy-
matous cortex becomes very thick-walled, while the thin-
walled outer cortex shrivels, though it does not fall
altogether away, as in other Dicræas to be considered
below. The lower part of the vascular bundle now also
exhibits the characteristic xylem elements, in the form of
annular and spiral vessels, while the upper part is phloem
tissue. Until this time there has of course been no use for
water-conducting tissue or xylem of any kind, but now the
floriferous part of the thallus is to come above the water.
No ligniflcation, however, seems to take place, or at least, the
walls, though f thickened, give no reaction with aniline
chloride or phloroglucin. Beyond the floriferous part the
thickening of the walls of the cortical collenchyma and of
the vascular strand does not take place. ^
The formation of the floral shoots in this genus shows an
advance on that in the preceding genera with root-thalli, in
that the whole shoot is floral only, is very short, and bears
346 WILLIS : MORPHOLOGY OP THE PODOSTBMACEÆ
one flower instead of many. The axis, hitherto evanescent,
comes above the surface of the thallus, and developes more
leaves. These differ from the previous vegetative leaves, in
that the sheathing base is larger and the blade smaller. The
axis rarely exceeds a few millimetres in length, exclusive
of the flower-stalk. The tips of the leaves shrivel away,
and very soon only the sheathing lower parts are left.
There are usually about four of these “ bracts ” below the
flower, the two upper being larger than the two lower ;
their shape is usually long and narrow, but sometimes shows
an approximation to the cowl or helmet form characteristic
of D. stylosa, &c. The upper side of the bract is thicker
than the lower, just as in the unaltered leaves of Podoste-
mon subulatus. At the end of the axis is the solitary flower,
enclosed as yet in its spathe, the latter being very like that
of Podostemon. PI. XYIII., flg. 11, shows a section at this
period through the thallus and secondary shoot, and shows
the shrivelled cortex and the vascular bundle branching to
the secondary shoot.
The spathe remains closed until exposure to the air, with
the flower inside it quite ready for expansion. Presently,
with the stiffening of the tissues the flowering part of the
thallus becomes fairly rigid, only the non-floriferous tip still
undulating in the water ; very often the flowering part
becomes more or less erect upon the rock. Soon after, with
the continual fall of the water-level, the flower-spathes
begin to be exposed to the air, and as this happens they split
and the flowers emerge. Thus, as all the flowers are ready
for immediate expansion, it happens that the actual order
in which they ultimately open is determined, not by any
morphological construction of the plant itself, but by the
orientation of the floriferous thallus with regard to the
water-level. As very often the thallus is nearly erect, it
happens that the order of opening is a descending one, so
commonly in fact that this plant is often quoted in the text
books as one in which an apparent raceme is really centri-
fugal. If the thallus happen to lie out horizontally, and be
OF CEYLON AND INDIA.
347
a little twisted, so that one side is nearer the surface of the
water than the other, then the flowers on this side will all
open before those on the other side, and may even have
ripened their fruits, while the flowers on the under side of
the thallus are still in the bud.
The flower is described and flgured by Warming ; in
general it is very like that of Podostemon subulatus, already
described and flgured above. It appears to be chiefly self-
pollinated, like the other Ceylon forms, and practically all
flowers set seed freely. The pedicel has the usual deciduous
cortex.
The fruit is very similar in structure to that of Podoste-
mon, but has equal lobes. PL XVIII., flg. 12, shows the ovary
in cross section ; it has the same wall structure and other
features as in Podostemon, already considered, but on
looking at the fruit in PI. XX., fig. 2, the essential difference
between the two genera can be easily seen. In Dicræa the
ribs are symmetrical on both lobes of the fruit, and from
each lobe three ribs run down into the pedicel. The
dehiscence rib is also symmetrically arranged, and the fruit
splits accurately down the middle line into two equal lobes,
which persists for some time on the stalk. As soon as
exposed to the air, the non-floriferous parts of the thalli
usually wither up and drop off, leaving the short woody
fruiting stalks shown on the left hand in PI. XIX.
Rejuvenescence.—Lil^Q the rest this species has a large
capacity for regeneration of injured growing points and the
production of new ones from almost any part of the thallus
so long as submerged. The creeping thallus is able to stand
considerable exposure without dying, and may revive and
form new growing points. The process of regeneration of
broken or injured tips has been described by Warming ; the
new tip forms in the thallus a little way back from the
injury, and growth is resumed. The creeping thallus also
very readily gives rise to new growing points by endogenous
(47)
MS WILLIS : MOilPHOLOGY OF THE PODOSTEMACBÆ
formation ; these grow along horizontally for a short distance,
and then give rise to vertical branches, exactly as the
primary thallus does. PL XX., fig. 1, shows a piece of reju-
venescent thallus from a plant which had been submerged by
a rise of water in the fiowering season before it was actually
dead. As the water falls in the dry weather, the lower parts
of the plants usually exhibit a large amount of this sort of
rejuvenescence, and towards the end of the dry weather
there is usually quite a large crop of plants with floating
roots from 5 to 8 inches long growing upon the rocks. These
roots show no sign of transforming their leafy shoots to
floral shoots the same year.
This species is well represented in the herbaria, but the
leaves of the shoots at the outer end of the thallus are very
brittle when dry, and have very often disappeared, some-
times even the whole of the foliiferous portion of the thallus
is missing. When this is the case it is perhaps usually due
to the fact that the specimens were gathered when the fruit
was ripe, by which time the terminal portion of the thallus
has usually disappeared, having withered up and broken oft'.
The leafy tips of the bracts, too, fall away to a large extent.
The size of all parts varies greatly. E.g.y the pedicel of the ripe
fruit, described as 1/5 inch long by Trimen, 6-8 mm. by
Weddell, 8 mm. by Tulasne, varies in the specimens in the
Peradeniya herbarium from 3 to 9 mm,; perhaps 6-7 is the
most usual length. The flower bud varies much in bulk,
and also the fruit, which varies from a length and breadth
of 2*2 and 1*25 mm. to 1*5 and *75 mm.; the largest fruits
are thus four times the bulk of the small. Some specimens
recently found in the Bambarabotuwa river, a tributary of
the Kalu-ganga, in Ceylon, by Mr. H. F. Macmillan, show a
smaller habit altogether. The floral shoots in many of the
specimens are only monostichous. The bracts also vary in
the direction of the more cowl-shaped form of those of P.
algæformis.
OP CEYLON AND INDIA.
349
Dicræa dichotonna, Tul.
(Plates XIV., XX.)
This species, as explained in the preceding paper, includes
the four forms, D. dichotoma, Wightii, rigida, longifolia,
originally described by Gardner and Wight, but which I do
not think are deserving of even varietal rank, depending
apparently on differences due to the different depths from
which the material was taken. I have carefully examined
the herbarium material, and have also collected for myself
in the Paikara river, and though further examination on the
spot is required, I do not think there is any good evidence
at present for separation of the forms named.
Habitat. — Like D. olongata, this species grows in rather
rapid and broken water. In the Paikara river it is frequently
mixed with Hydrobryum olivaceum, var. griseum, the only
other form that occurs in that locality.
Dry Season Appearance. — A stone covered with fruiting
material is shown in PI. XIV. (top, right). Brown withered
thalli about 5-8 cm. long, thin, ribbondike, more or less
zigzag, may be seen attached to the rocks for at least the
greater part of their length, though the ends are usually
more or less free. Standing vertically upon these are the
fruiting shoots, very like those of D. elongata, in fact it is
not always easy to tell which is which when, as sometimes
happens, only the lower part of the thallus of the latter
bears fruits. Under water plants may be found in January
at any rate, which illustrate the mature structure sufficiently
well, but it is desirable that the life-history should be fully
worked out.
Mature Structure.— seen at the end of December, the
thalli are about 10-15 cm. long or more, lying very closely
down upon the rock, and with much less marked difference
between the creeping and drifting parts than in D. elongata.
Usually the basal 3-5 cm. are attached to the rock, and the
rest free, but in a line with the creeping part, and often
attached at one or two points by haptera. The haptera
350 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
are often very large — to 1 cm. long — and much branched in
this species. The drifting part is usually fairly frequently
branched, the branches also drifting and repeating the
structure of the original thallus ; the creeping part is less
often branched.
The thallus is flattened, but not so much so as in the
species to be considered below ; at the base it may be as
much as 4 mm. wide and 3 mm. thick, and tapers to the tip.
It is not unlike the thallus of Podostemon Barberi, but longer
and narrower, and in the older parts it grows in thickness
like the thallus of D. elongata, more especially in the central
part, while the marginal portions shrivel (PI. XX., fig. 4),
The vascular strand and the general anatomy of the thallus
are much as in the preceding species. The secondary shoots
are about the same distance apart, and the thallus is often
more or less zigzag, small lateral projections being formed
under each secondary shoot (PL XX., figs. 3, 5).
The general structure and history of the secondary shoots
is the same in all essential features as in D. elongata, and the
basal third or even half, or more, of the thallus becomes
ultimately floriferous. The formation of the bracts by the
metamorphosis of the leaves is clearer in this species than
in the last, and requires a little explanation, as the method of
formation is common to other genera of the order, and the
examination of herbarium material in various stages of the
metamorphosis has led to many errors and much confusion.
The figures in PL XX. show the process clearly. In fig. 6
the secondary shoot is beginning to develop a flower ; it
still has several long leaves with sheathing bases. The bases
enlarge, especially those of the two uppermost leaves, and
the tips of the leaves begin to shrivel as in fig. 7. Then
follows the stage shown in fig. 8, where the tips have mostly
fallen away, and finally the process reaches the stage shown
in fig. 9, where the flower is open, with its spathe split into
several teeth at the tip ; at the base of the spathe are two large
more or less cowl-shaped bracts with no tips, and below these
two or more leaves, which are only little enlarged, but
OF CEYLON AND INDIA.
35]
which are usually considerably stretched by the growth
of the outermost bracts and the flower. Many specific
characterizations have been chiefly based on differences
in the form of the tips of the bracts, while in reality these
differences simply express different stages of the process
above explained, and until it is complete it is not safe
to base any diagnosis on the form of the tip or keel of the
bracts.
As mentioned in the preceding paper, several species have
been made from the Paikara material, the separation char-
acters chosen being chiefly the length and number of the
bracts ; D. Wightii, for instance, depends for its character
simply on the fact of its having been taken from deeper
water, before the long tips of the leaves had fallen. I
examined all the plants I could at Paikara, and came to the
conclusion that the forms mentioned could not be regarded
as anything but phases.
The flower of this species is very similar In all points to
that of D. elongata, and the fruit is the same, but with
broader dehiscence ribs (PI. XX., fig. 10).
Rejuvenescence by the formation of new growing points
on broken or resubmerged thalli appears to take place to a
considerable extent.
The herbarium material of D. dichotoma is on the whole
good, apart from the confusion caused over the development
of the bracts, which has led to its being divided into many
species.
Dicræa WaiSichii, Tul.
(Plates XX.-XXI.)
As explained in the preceding paper, I include Weddell’s
D. pterophylla in this species, it having been separated from
it simply on account of misconception of the process of bract
development ; Weddell examined spirit material collected
by Sir J. D. Hooker, while Tulasne and Griffith used dry
material, in which the keel \vas more disintegrated. I divide
the species into two varieties : of the second, D. Wallichii
352 WILLIS: MORPHOLOGY OP THE PODOSTEMACEÆ
striata, the Banna form, I have only seen herbarium material,
and probably enough when good material is available it
may prove specifically distinct. The following remarks
refer only to the Khasia mountain form, D. Wallichii
Khasiana, which I have myself studied at Cherrapunji, the
district in which the plant was originally collected. I have
also examined the herbarium material.
Habitat. —'Vh.Q plant grows in very similar situations to
D. elongata, on rocky substratum in rapidly flowing water.
Owing to the enormous rainfall of Cherrapunji, there must
be considerable depth and force of water during the vege-
tative season, I found it in one or two places growing
slightly intermingled with Hydrobryum lichenoides.
Dry Season Appearance. — A number of specimens are
figured in PL XXL As usually seen at Cherra, the plant
consists at this time of brown thalli, a few millimetres wide
and almost as thick, attached at the base by a sort of foot,
and running outwards from it, usually down stream, in a
branched laciniate form, to perhaps 5-10 cm. long. On the
ends of the laciniate arms the fruits are borne, usually one
on each, on a stalk of about 8 mm. long.
Mature Structure. — The most of the dried material in the
herbaria gives little clue to the mature structure of the living
plant, which is really very like that of D. stylosa, com-
paratively little branched, and with broad lobes, not with
laciniæ. In most essential points the structure is that of the
latter species, to wiiich reference may be made for details.
The thallus is usually attached at the base only, or at the
base and a few outer points, only rarely at all points. At
the base, where the primary axis formerly stood, there is a
sort of stout cup-shaped foot, seen from below in PL XX.,
fig. 14. From this the thallus spreads out, chiefly in one
direction, that of the prevailing current, so far as I have been
able to observe. In general it has the habit and construction
of a Fucus, with broad crisp undulated lobes just as
described below for D. stylosa, var. fucoides. Plate XXI.
shows a series of specimens from above and below the water.
OF CEYLON AND INDIA.
353
and with the explanation that accompanies it is sufficient to
clear up the somewhat confusing appearance of the various
existing specimens of this species and its synonym.
The thallus grows by a broad growing point like that
described below in D. stylosa fucoides, with exogenous
branching; it forms endogenous secondary shoots on the
upper edges, to which run branches of the Central vascular
bundle (PI. XX., fig. 11). Probably when in full vegetative
growth in September the plant is very like D. stylosa
fucoides as shown in PI. XXII., fig. 3, but towards the end of
the season the tips of the thallus lobes are commonly found
to have fallen away or to be breaking off, so that the general
form is commonly like that shown in fig. 12, PI. XX. In
this specimen the vascular bundles are shown, leading to
the marginal secondary shoots (themselves not shown).
Ultimately, as usual, some but rarely all of the secondary
shoots become floriferous, and here a peculiarity of this
species shows itself, in that instead of the flowers being
concentrated on one part of the thallus, they are scattered
over the whole or most of it, in each lobe some of the shoots
becoming floriferous, the rest remaining sterile. Round the
vascular bundles leading to the developing flower the tissue
as usual becomes thick-walled and brown, while the bundles
leading to the sterile shoots are not so surrounded. The
former bundles show clearly as brown bands through the
living thallus fig. 13). As the flowers develop the inter-
mediate tissues become disorganized and begin to break
away, so that finally nothing is left but the brown bands of
thick-walled tissue containing the vascular bundles leading
to the flowers. Thus the remarkable differences seen in the
different collected material are easily explained. The
process is not usually complete till the plant is fully exposed
on the rock, and I am inclined to think, but cannot be sure,
that the flowers emerge from the water as soon as it gets
shallow. Griffith’s material mostly shows the thallus with
the thin marginal parts between the bundles still present,
though the fruits are ripe. Other material has been
354 WILLIS : MORPHOLOGY OF THE PODOSTEMACBÆ
gathered from dry rocks, and has completely fimbriate
thalli without thin margins. Others again have been
collected quite early, and show the process only in its early
stages, and so on.
The floral shoot develops from the vegetative, as in
D. dichotoma or elongata, but does not become so much
lengthened. The basal sheathing parts of the leaves, which
form the scaly bracts, are proportionately much shorter and
broader, helmet-shaped in fact. The spathe does not show
at first between the two outer bracts, and the floral bud is
consequently as broad as long and closely sessile on the
thallus. The open flower itself (PI. XX., fig. 15) is very
similar to that of the preceding species.
Rejuvenescence. — I found this process going on at Cherra-
punji in many plants which still remained submerged ; the
process was exactly like that in D. stylosa, var. fucoides, to
be described below.
Dicræa stylosa, Wight.
(Plates XXII.— XXIV.)
As explained in the preceding paper, I include under this
name several Ceylon and South Indian forms, which will
probably ultimately prove to be specifically distinct, but to
which, pending detailed knowledge of all forms from many
localities, I have only given varietal rank. We shall deal
with the forms in order. Warming has described the Ceylon
form D. stylosa fucoides from Hakinda dry season material
sent by Trimen, and there is comparatively little to add to
his description, except for the earlier stages of the life-
history.
D, stylosa laciniata, Willis.
HaMtat.--ThX^ form is common in the Kandy District,
Ceylon ; it is often confused with D. elongata, to which it
has considerable resemblance, and with which it is often
mixed. It also occurs with D. stylosa fucoides, and less often
with Podostemon subulatus or Hydrobryum olivaceum, very
OP CEYLON AND INDIA.
):);■) 5
rarel}^ with Lawia zej- lanica. It grows iu rapid water which
Avhen shallow is liable to be broken, and lives at a rather
greater average depth than some ot the other forms.
Dry Season A'lipearance. — In general this is almost
exactly like that of D. dichotoma, flat, narrow, prostrate
thalli, with erect fruit stalks.
Life History and Mature Structure, — In consequence of
the infrequent flowering of this species, and the somewhat
awkward places which it mostly affects at Hakinda, I have
not been able to obtain seedlings ; probably the early stages
are like those of D. elongata, which the plant resembles very
closely in habit. When fully grown it is about the same size
as D. elongata, with prostrate creeping thalli upon the rocks,
giving off long drifting thalli, often as much as 50 cm, long,
and sparingly branched. The chief distinction in habit from
D. elongata is that the thalli are flattened, like very narrow
ribbons, just like those of D. dichotoma ; they are, how-
ever. much longer than those of the latter species.
The thallus has a growing point like that of D. elongata,
but flattened and broader (PI. XXII., fig. 1), with the
secondary shoots developing on the upper 'sides of the mar-
gins in the usual way. The upper side, near to the tip, is
very convex, the lower approximately flat with a slight
convexity over the central part, in which the vascular strand
is situated. The internal structure is much like that of the
species already described. Subsequently the thallus grows
in thickness, much as in var. fucoides, described below, by
meristematic activity in the central part above, and still more
below, the vascular bundle, while the marginal portions do
not grow to correspond. In this way the centre becomes a
stout rib, very prominent on the lower side. The vascular
bundle is like that of D. elongata, but sometimes it is divided
into two by a band of collenchymatous cells at right angles
to the substratum. The thallus grows rapidly in length,
and occasionally branches exogenously, the branches lying
parallel to the thallus from which they sprang.
(48J
;556 WTLLTS : MORPHOLOGY OF THE PODOSTEMACEÆ
The secondar}^ shoots are like those of the other Dicræas,
and ultimately the lower end of the thallns, rarely more than
about one-third of it, becomes floi-iferoiis, just as in D. elon-
gata. The floral shoots form in the way described above
for D. dichotoma, but the bracts are broader in proportion
to their length, just as in D. Wallichii, and when the tips
have fallen away, are almost helmet-shaped. The flower
emerges as soon as the tip of the spathe is exposed to the air,
and is wind-fertilized, probably, as usual, mainly by its own
pollen. The non-flowering part of the thallns does not
become woody, and falls away on exposure, just as in
D. elongata.
Rejuvenescence takes place just as in D. elongata, and is
very common.
1). sty losa fnco ides, Willis.
Habitat. — -In general this is the same as that affected by
the preceding form, with which it is commonly found
associated ; it is less often found mixed with D. elongata,
Lawia zeylanica, or Hydrobryum olivaceum. Both this
form and the last may be seen in perfection, forming a very
striking sight, at the confluence of the (Juru-oya with the
Hulu-ganga, near Teldeniya, a little below the ford over the
former ; the water is clear, not muddy like that of the
Mahaweli-ganga, and the plants can be seen to some depth.
Dry Season Appearance. — When in fruit the thalli are
almost exactly like those of the preceding, but a trifle broader.
Germination and Life History. — The seedlings are very
hard to find, and I have only been able to get two of them,
both at the stage shown in PI. XXII., fig. 2. The primary
axis is like that of D. elongata, short and stout, with a hap-
terous base and endogenous thallns arising nearly opposite
the hapteron. The growing apex is broken off in both my
specimens, but is probably like that of the mature plant.
The thallns broadens out at once much more than in D. elon-
gata, rapidly becoming nearly circular in outline, and is
OF CEYLON AND INDIA.
857
slightly upcnrved at the edges. On the lower side haptera
and rhizoids are formed in the usual way. The next stage
is seen in August and September, Avhen plants are found
like that shown in PI. XXII., fig. 3, a few inches across,
starting at the base from a stout hapterous foot like tliat of
D. Wallichii, and shown in section in fig. 4. From this the
thaPus grows outwards and usually slightly upwards, and is
perhaps most often only attached at the foot, but very com-
monly also at one or more outer points, usually by haptera,
but sometimes it is found touching the rock and fastened by
root-hairs. Sometimes the thallus is entirely prostrate and
creeping, attached to the rock at all points, and then looks
just like that of Griffithella, shown in PI. XXV., fig. 1. I
have not been able to determine satisfactorily what causes
one form or habit to be assumed rather than another, but it
seems fairly clear that the fully creeping form is rare, except
when there is ample space available. When the plants are
crowded they tend to grow more upwards, attached only by
the feet. When they grow near the edge of a rock the plants
tend to give off streaming thalli like those of D. stylosa
laciniata, but broader. When fully grown the plant may
reach 20 cm. or more, with the thalli frequently branched or
lobed ; the lobes are exogenously formed and are often
as much as 7-10 mm. wide. At the same time they are also
commonly more or less undulate in form, with margins
curved up and down, and the whole plant, Avith its red or
olive colour, has an extraordinary resemblance to a
Fucus or similar alga, especially Avhen a large mass of
the form attached only by the foot is found groAving
together.
The groAving point is rounded and rather upcnrved at the
tip, on Avhich is a root-cap almost completely on the upper
side, as already described by Warming for this species (cf.
PI. XXII., fig. 6). As seen in section (fig. 5) this develops
from the meristem. at the apex in the usual Avay, and the
same meristem. gives rise to the general tissues of the thallus.
Details need not be given here. At the sides of the groAving
Î358 WILLIS: MORPHOLOGY OP THS PODOSTEMACEÆ
apex, which very rapidly widens to its full size, the endo-
genous secondary shoots may be seen developing, with a
branch of the central vascular strand going to each. The
apex is continually branching, so that the thallus assumes
the general form shown in PL XXII., fig. 3.
In cross section the thallus when in young vegetative con-
dition shows a slightly thickened rib in the middle, and
long thin lateral wings turned down at the extreme margin
(PI. XXII., fig. 7). The vascular strand is rather below the
centre, as usual. It is shown on a larger scale in fig. 8, and
has two well-marked groups of tissue, as described by
Warming, more or less separated by prosenchymatous tissue
between. Its subsequent growth, and the meristematic
activity that goes on, are closely similar to what occurs in
D. elongata, and have been described in part by Warming,
whose accounts of the peculiar bundles of the very dorsi ven-
tral types of Podostemaceæ appear to have been somewhat
overlooked by anatomists.
Growth in thickness also goes on in the thallus itself out-
side the bundle. Instead of being, as in D. elongata, almost
equal throughout the thallus tissues, it is practically confined
to the tissue above and beloAV the bundles, and in these there
are more tangential than vertical divisions, especially on the
upper side. PL XXII., fig. 9, shows a transverse section at
an early stage, a little to the right of the bundle. The upper
cells are mainly dividing tangentially, while the lower are
also dividing vertically, the result being an upeurving of the
thallus. It is in this way that the shape of the foot arises
and similar causes produce the curvatures and irregularities
which are so marked in these Fucus-like thalli. The lateral
wings of the thallus do not take part in this growth in
thickness, and thus a stout central rib is gradually formed,
just as in Fucus itself. In transverse section near the base
of an old thallus the result of this continual tangential
division is often very striking, as shown in Warming’s figure
(42, II., PI. XI., fig. 22), where there are long, vertical,
parallel rows of cells not unlike a very deep palisade layer
OF CEYLON AND INDIA.
359
of a leaf. Whether this development has any functional
object like that of the palisade layer must be left for future
determination ; it seems a little improbable in view of the
fact that the younger parts of the thallns, which probably
do most of the assimilatory work, are without it. As the
development proceeds, the thin marginal parts of the thallns
commonly shrivel up, and often fall away altogether, leaving
only the stout rib, so that a piece of the base of the thallns
of this form is indistinguishable from that of var.
laciniata, though the early stages of the two are so utterly
different.
The .secondary shoots in the vegetative condition are just
like those of the other Dicræas (PI. XXII., fig. 3). Ulti-
mately the lower ones, usually for about a quarter of the
length of the thalius, but sometimes more, become floriferous
in the usual way by growing out from the thalius, enlarging
the sheaths of a few leaves and losing their tips, thus giving
rise to broadly helmet-shaped bracts. The spathe is en-
closed in the upper bracts till a later period than in D.
elongata, and frequently the bracts do not even separate till
the air touches them. Although the vegetative shoots are
erect, the flowering ones usually lie more or less prostrate
until the curving upwards of the spathe is accompanied by
more or less curvature of the bracts. The spathe is broadly
funnel-shaped (PL XXIV., fig. 1). The flower is of the usual
Dicræa type, on a short stalk, with short stigmas and stamens.
It is wind-fertilized and usually autogamous, and practically
all flowers appear to set seed. This form at Hakinda, how-
ever, lives so deeply submerged that it does not always set
very many flowers. The fruit stands on a pedicel about 6
mm, long which becomes thin and elastic in the usual way by
dropping its cortical tissues.
Rejuvenescence.— species exhibits very extensive pro-
duction of new growdng points in case of any injury to the
thallns, such as the breaking off of a portion of it, or injury
to the tip caused by exposure to air. New growing points
3 60 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
form in the usual way, endogenously, behind the old one, or
perhaps anywhere in the tissue, and soon grow out into new
thalli. The older basal parts of the thalli can survive a long
exposure, and may be found rejuvenescing at any time of
temporary rise of the water in the dry season.
D. sUjlosa Boiirdillonii, Willis.
This form, which is the original type of the species, is
common in herbaria. My spirit material was collected at
Mundakayam in Travancore by Mr. Bourdillon.
The dry season appearance is very similar to that of the
last form, but the plant is larger in all respects, and much
longer, while the non-floriferous parts of the thalli not un-
commonly seem to persist, though they lose their leaves.
The thallus is large, often to as much as 30 cm. long, broadly
ribbon-like, and much branched, the width of the branches
decreasing towards the outer ends. It is attached by a foot,
and very often by one or more outer points. The lower
part becomes floriferous as usual, with closely crowded
shoots, just as in fucoides. The vascular bundles show large
and brown where they lead to flowers, just as in D. Wallichii,
and the marginal parts of the thalli break away more
decidedly than in fucoides ; they often show thin and
membranous in herbarium specimens.
The flowers (PI. XXIV., fig. 2) are as in the preceding
form, but larger, with longer stamens and with stigmas
often as long as the ovary, but varying considerably in
length. The stigmatic character is that on which Wight
based the species, but is too inconstant tobe taken by itself.
The fruit is borne on a long stalk with the usual deciduous
cortex.
D. stylosa algceformis^ Willis.
This is Beddome’s D. algæformis (not Trimen’s), from
the Anamalais, figured by him (3), but in some points
OF CEYLON AND INDIA.
361
incorrectly, as explained in my first paper. I obtained
material in the Anamalais, where it grew in a rapid stream,
sharing its habitat only with Hydrobryum lichenoides. It
agrees in all essential points with the form last described,
but has shorter stigmas on the whole, and the stamens do
not greatly exceed the ovary as in that form. In some of
Mr. Barber’s material, parts of the thallus are very broad (to
2 cm.), and there are a considerable number of secondary
shoots on the upper surface as well as at the margins. This
simply implies, of course, a greater proportionate growth of
the marginal parts, as compared with the central part and
vascular bundles.
D. stylosa kanarensis, Willis.
The material which I have examined was collected by
Mr. Barber in rapids in S. Kanara, where its habitat
was shared with Griffithella Hookeriana and Lawia
zeylanica.
The habit of the plant is practically the same as that of
the form fucoides, and the plants are of about the same
size. Two chief points of difference must be noted. As
in D. Wallichii, the floriferous secondary shoots are not all
together, but are divided from one another by non-floriferous
(PI. XXIV., fig. 3). The vascular bundles leading to the
floriferous shoots become surrounded by the usual dark
brown woody layer, while the others remain thin and indis-
tinct. Presumably the thin parts of the thalli ultimately
shrivel and perhaps drop off, but I have no material collected
at a late enough stage to be sure of this. The bracts are also
peculiar ; when young they are long and rather incurved, and
they develop the cowl-like sheathing bases in the usual way.
The tips, however, do not break off or shrivel close to the
sheath, as in other Dicræas, but some distance above, so that
a more or less persistent acuminate tip is left, even after
drying. The fruit commonly stands on a very long pedicel,
but sometimes on a much shorter one.
WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
Oioræa s^inor^ Wedd.
(Plate XXIV.)
As mentioned in the preceding paper, the autonomy of
the species is very doubtful, and further investigation on
the spot where it grows is required. Griffith’s dry material
looks distinct, on account of the very short fruit stalk (PI.
XXIY., fig. 6), but the spirit material collected at Nongkhlaw
by Sir J. D. Hooker, and named by Weddell, looks to me
very like D. Wallichii, though of course, having been collect-
ed in a different locality from the typical D. Wallichii, it
will probably prove different in detail. A specimen is
figured in PI. XXIY., fig. 4, and a flower bud (secondary
shoot) in fig. 5. In view of this uncertainty there is no
need to describe it further.
Comparing Dicræa with Tristicha and Podostemon, it is
evident that at least one line of evolution which was pointed
at by a comparison of those genera between themselves has
here been carried out much further, viz., the dwarfing and
reduction of the size of the individual secondary shoots. In
Dicræa they are very small, with no appreciable axis until
they flower, and the number of fl.owers on each is still further
reduced than even in Podostemon, each shoot bearing one
only. The number of shoots, on the other hand, is so greatly
increased that, though only part of them flower, the number
of flowers is probably about as large as in Tristicha or
Podostemon. The advantages gained by the reduction of
the size of the shoots are evident ; the plant is able to live in
shallower water, and it suffers less from exposure resulting
in the death of the exposed shoots ; instead of losing a large
and complex shoot as happens in Tristicha and to a less extent
in Podostemon, it loses only a few leaves, and the small
portion of thallus on which they stand or which may be
exposed ; at the same time this thallus retains its vitality
for a considerable period, and may revive and form new
OF CEYLON AND INDIA.
363
growing points if again submerged. On the other hand, the
reduction of the shoots to little tufts of leaves is attended
with a corresponding loss of assimilatory capacity, but this
is made up by the remarkable development of the thallus
into a leaf-like structure, so that the mature plant is not
unlike the large leafy Podostemaceæ of South America,
though in the latter the leaves are of the ordinary kind.
The thallus of Dicræa is still pretty evidently homologous
with the root-thalli of the preceding genera, but regarded
simply as a ‘‘root,” it is one of the most remarkable struc-
tures that can be imagined under that category— endogenously
developed from the main primary axis, it is true, and with a
semblance of a root-cap, but exogenously branched, usually
not attached to the rock, except at the base and perhaps else-
where, but drifting freely in the water, and performing the
chief functions of assimilation, besides bearing the florif erous
shoots.
Probably the drifting form of thallus should be regarded
as derived from the creeping, and there is evidently as yet
no very marked separation of the two types, except in D.
elongata ; in the more algal types of Dicræa there is a good
deal of polymorphism, and the thallus may assume almost
any form, whether creeping or drifting. D. stylosa fucoides
in particular is of interest in this way, and shows the first
signs of the remarkable polymorphism which we shall see
carried to extreme in Griffithella.
The dorsiventrality of Dicræa is extreme, and shows itself
in many features, such as the growth in thickness of the
thallus, the structure of the vascular tissue, &c. It is notice-
able, however, that it is not appreciably exhibited in the
secondary shoots in their vegetative stage, though it is shown
in the corresponding stages of Tristicha and Podostemon.
Here we have perhaps a case in which the evolution has
been through dorsiventrality back to radial symmetry. As
soon as the secondary shoots become fiorif erous, they become
dorsiventral once more, though not so markedly so as those
(49)
364 WILLIS : MORPHOLOGY OF THE PODOSTEMACBÆ
of Hydrobryum, &c. The flower itself, with its symmetrical
ovary, is less dorsi ventral than that of Podostemon, and
conies nearer to the floral types of the less highly modified
groups of Podostemaceæ.
GRIFFITHELLA.
[Warming ; Willis, Rev. Podost. Ind., Ann. Perad., I., p. 231.]
Under this genus I include the curious species first found
by Law, described as a Mniopsis by Tulasne on account of its
smooth fruit, and placed in Podostemon by Weddell, Ben-
tham, and others. On the whole, perhaps, as explained in
the preceding paper, it comes nearest to Mniopsis, but on
account of its peculiar thallus morphology, its reduced
secondary shoots, and the different stigmas, I am inclined,
though with some diffidence, to regard it for the present as
generically distinct. Warming (42, VI., p. 13) has described
a second species, G-. Willisiana, from herbarium material, but
I am hardly inclined to regard this as a separate species
until further material has been examined from Kanara, and
I have therefore included it as a variety in G. Hookeriana.
The genus also seems, as Warming points out, to have
relationships to the Javanese Cladopus,
In general we may say that the genus, as at present known,
has the morphology of Dicræa in the vegetative parts, and
of Mniopsis in the floral.
GriffStheBla Hookeriana, Warming.
(Plates XXIV.-XXVI.)
For material of this extremely interesting plant I am
indebted to my friend Mr. Barber, who collected the
remarkable series of forms figured in Plates XXV. and
XXVI. (left), and to Mr. R. K. Bhide, who collected some
(Plate XXVI., right) at Atgaon, west of Poona. I sought for
it in vain on my own visit to the Bombay Ghats.
OP CEYLON AND INDIA.
365
Habitat. — The plant appears to grow in moderately rapid
water, such as is affected by the Dicræas. I have found it
mixed in my material with Lawia zeylanica, Podostemon
Barberi, Dicræa stylosa kanarensis, and Tristicha ramosis-
sima.
Dry Season Appearance. — The general appearance of the
plants at the end of the year is shown in the three smaller
rocks in Plate XXV., where also may be seen the first
indication of the very remarkable polymorphism, which is
characteristic of this species. On the large lower stone is a
large plant just exposed and flowering. This has a closely
attached creeping branched thallus, with marginal flowers.
A similar and larger plant, still in the vegetative condition,
is shown on tbe large rock figured on the left. On the
smaller stones above are thalli in fruit, evidently, though
shrivelled, of the most various shapes, and much smaller
than the creeping one below. Examination with a lens
shows that some of them are flat upon the rock and branched,
others cup-like, some like discs or cups on stalks, &c. The
fruit is smooth, almost spherical, with one smaller valve
splitting off obliquely from a larger persistent one. The
seeds are shed upon the rocks in the usual way, and the
early stages of the life are no doubt similar to those in the
rest of the group. It is very much to be desired that this
plant should be followed throughout its life- history. The
material at my disposal was all collected at the end of the
life, and contains no stages that show the primary axis or
the development of the peculiar forms assumed by the
thalli.
Mature Structure. — The chief interest in this species
centres in the thallus and its extraordinary polymorphism.
In general it has the morphology of that of Dicræa. The
growing point in all my specimens was more or less shrivelled,
and I was not able to definitely make out its construction,
but on the whole it seems to be like that of Dicræa stylosa,
var. fucoides, broadly obtuse in shape ; whether there is a
properly developed root-cap or not I was not able to decide*
366 WILLIS : MORPHOLOGY OP THE PODOSTEMACEÆ
The thalliis in the creeping form, with which we shall
first deal, is like that of Dicræa stylosa in transverse section,
and grows in the same way by tangential divisions of its
cortical cells. It may be as much as 1 cm. wide, and creeps
along the rock, closely attached to it by root-hairs and
occasional hapterous outgrowths, branching exogenously,
till at last a large plant like that figured in the left hand of
PI. XXV. may arise. At intervals of a few millimetres
along the upper side of the margins the secondary shoots
arise as seen in the figures, and endogenously as usual ;
their first appearance is extremely close to the growing apex.
These shoots are very like those of a Dicræa, but are often
rather more prostrate, with distichous sheathing leaves to
about 5 mm. long. As the figure indicates, these leaves
probably only perform a small portion of the work of
assimilation, being small, as in Dicræa, in proportion to the
area of the thallus. Vascular bundles lead from the main
bundle of the thallus to the leafy shoots, in the ordinary
way, and the thallus grows in thickness like that of Dicræa
by tangential division of the cortex.
We may now go on to deal with the extraordinary series
of forms shown in Pis. XXV, and XXVI. All in XXV. and
the left-hand half of XXVI. are the var. Willisiana collected
by Mr. Barber, while the rest are some forms collected near
Atgaon. These are not the only Atgaon forms, but are those
not well represented in the Kanara material. The cup and
disc forms are frequent at Atgaon, but not the simple
creeping form.
Not infrequently, as in some of the plants on XXV., 2, and
in XXVI., 5, 6, the creeping form, instead of growing to a
large size, remains short,* and more or less discoid orlobed.
When the growth of the actual apices is not much faster
than that of the tissue between, the form approximates to a
* I am inclined to think that the large creeping form may perhaps be
varietally distinct from the small one, though they agree very closely
indeed in all other points than size ; the question can only be settled by
detailed study on the spot where they grow.
OF CEYLON AND INDIA.
367
disc, but when the apical growth is rapid, it is lobed oi
branched. When the disc form grows old, it is usually
stout, and with the margins slightly turned up. The
upward curve seems to be brought about as in Dicræa
stylosa fucoides by greater meristematic activity and growth
near to the edges.
Accentuation of the growth of the sides of the base of the
thallus results in a form like a bowl or cup, lobed or not,
according to the relative rate of growth of the apices. When
once the corner is turned, so to speak, the thallus continues
to grow away from the rock in an ascending direction, and
so the cup may reach a considerable size, as in several of
the examples figured in PI. XXVI.
The simplicity of the cup form may be complicated by
further irregularity of growth in length and thickness at
different parts, resulting in the very involved forms seen
on the right in PI. XXVI. In all but the last two of these
the foot or disc from which the plant starts may be seen.
If the development of the thalli be very uniform, and the
growing points do not outrun the tissue between, the cup or
bowl may be very symmetrical, as in XXVI., 1 and 4.
Sometimes one side grows faster than the other (this is the
case especially when the plant is on a sloping rock, the lower
side growing the faster), and a scuttle-shaped thallus is
formed like that in PI. XXVI., fig. 3.
Not infrequently the foot grows vertically by division
and elongation of its cells, and thus forms a shorter or
longer solid stalk (narrower than the actual foot, which
usually forms a disc like that of Tristicha ramosissima),
carrying up the disc or bowl upon its summit, as in
PI. XXVI., figs. 1, 2, 3, 10. The first-mentioned figure re-
presents a most symmetrical stalked cup, like the basal cup
of the alga Himanthalia lorea. It is particularly interesting
to see in these plants how many of the forms of the algæ
of moving water they reproduce.
The remarkable polymorphism of the thalli in this plant
is thus very simply explained anatomically, and forms an
368 WILLIS : MORPHOLOGY OF THE PODOS'PEMACEÆ
extremely interesting illustration of how plastic the form of
the plant may be when once it has got over the difficulties
presented in most plants above the mosses by the presence
of a more or less rigid skeletal tissue and by the absence of
merismatic activity in all parts. Here the form is deter-
mined before the bundles become stiffened in the flowering
season, and the capacity for renewed merismatic activity
possessed by nearly all the cells enables a great variety of
form to be produced, chiefly by simple irregularities or
difterences in the rate of growth of the different parts or
cells. Almost incredible though it seems at first glance, the
various forms in PI. XXV. and the left-hand part of PI. XXYI.
are all forms of one plant, not eTen varietally distinct
from one another, unless the large creeping form be distinct
from the small creeping one. Detailed work on the spot
with living plants is required to determine with certainty
whether there is any heredity of the form of any individual,
but so far as Mr. Barber’s observations go, and they confirm
what I have mentioned above as occurring in Dicræa stylosa
fucoides, the particular form of any individual seems
mainly, if not entirely, in a direct correlation with the
environment. Plants on the top of the rock, when crowded,
tend apparently to the nearly symmetrical cup or disc form,
and those on the sides of the rock to the form in which thie
lower side is produced to a greater length than the upper,
while when there is plenty of room the creeping form seems
more common. This last form, however, seems quite absent
in the Atgaon material.
The secondary shoots are very closely similar to those of
Dicræa, and some or all of them ultimately become flori-
ferous, while the tissues leading to them become woody in
the usual way. The bracts form just as in Dicræa, with
broad sheathing bases and deciduous tips (PI. XXIV., flg. 8).
The solitary terminal flower emerges from a spathe as usual,
and stands erect ; it is anemophilous, as in all the other
Indian forms. The general floral structure is like that of
Podostemon, with a dorsiventral ovary which ripens to
OP CEYLON AND INDIA.
369
anisolobous fruit with one deciduous valve. The stigmas
in the Atgaon form are generally subulate, but in the var.
Willisiana are very often ovate or almost cordate, notched
or even fimbriate or divided into two, like the stigmas of
the Hydrobryums to be considered below. The ovary wall
shows no vascular bundles, and the fruit consequently no
ribs (PI. XXIV., fig. 9), but, as fig. 10 shows, it has a stout
cutinized inner epidermis like that in the preceding genera,
then a layer of thick- wailed cells (f ), and outside that one
or two layers of sclerenchyma, followed by the outer thin-
walled epidermis and sometimes a thin-walled hypoderm.
The thin-walled cells fall away from the ripening fruit,
which has a smooth wall.
I have seen no evidence of rejuvenescence, but it probably
occurs as in Dicræa,
Comparing this genus with the preceding ones, it is
evident that it is very similar to Dicræa, so far as morpho-
logical adaptation to its mode of life is concerned ; it has
the same algoid form of thallus and reduced secondary
shoots. Its chief interest is in the great polymorphism,
probably the most extreme known among the higher plants,
and in the fact that this carries out to its highest develop-
ment what we saw begun in Dicræa. Strange though the
forms of the various thalli are, regarded only as structures
in flowering plants, they are parallel to the similar forms
that may be found among the algæ of moving water, just
as is the case with nearly all the other forms assumed by
the Podostemaceæ.
WILLISIA,
[Warming- ; Willis, Rev. Pod. Ind., Ann. Perad., I., p. 233.]
This very peculiar genus, so far as at present known, is
confined to the Anamalais and Burma ; I have collected
material in the former locality, and Mr. Barber has since
obtained some for me from the same place at an earlier time
370 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
of year. Willisia selaginoides, the species there found, was
first described and figured by Beddome (3), and has
hitherto remained almost unknown. It is a very remark-
able form, with no very evident near relationships among
the other Indian forms.
Willisia selaginoides, Wmg.
(Plates XXVIII.- XXX.)
Habitat. — I have only seen this plant at one place, in the
Sholai-aar near Monica estate in the Anamalais, at a height
of 3,500 feet above sea-level. It was growing in a steep rocky
gully, somewhat out of the main rush of the water, but
with a rapid current flowing in it. It was there accom-
panied, as PI. XXYIIL shows, by Hydrobryum lichenoides
Fentonii.
Dry Season Appearance. — This is sufficiently well shown
by Pis. XXVIIL, XXIX., and is very different from that of
all the other forms with which we have to deal. The rocks
are covered with little tufts of stems 2-7 cm. high, each tuft
often being composed of one plant only; they stand stiffly
erect, have four rows of scaly leaves each, and are termi-
nated each by a solitary fruit, typically half concealed
among the uppermost leaves, but often on a long stalk due
to the falling away of the tissues. Many stalks in the older
or longest exposed shoots have often lost the whole or nearly
the whole of their cortical tissues and their leaves in this
way. At the base of the shoots one can see indications of a
small thallus, on which they are closely crowded. The fruit
is smooth, with unequal valves, one remaining upon the
pedicel after the fall of the other. Usually at this period
some plants may also be found alive in the water, and it is
on material thus obtained that the description here given is
based. The shoots stand stiffly erect, and consequently the
flowers are exposed to the air, open, and ripen their fruits
long before the plant is actually killed by the exposure of
its thallus (of course, in all probability the thallus, as in the
OF CEYLON AND INDIA.
371
most of the Ceylon forms, probably lives for a considerable
period after being exposed, and probably may rejuvenesce).
The plant grows in a very rapid current at the place where
I studied it, and the erect shoots were in a state of constant
and rapid quivering movement to and fro.
Mature Structure. — The germination and the early stages
of the life-history of this plant should be of especial interest.
The earliest stages that are as yet known are those in the
material collected in November last by Mr. Barber, and
these are almost as mature as those collected by myself in
January, but show very well what I take to be the primary
axis. One of these specimens is figured in PI. XXX., fig. 2,
and shows at the base the clump of little floriferous shoots
just described, with two long shoots, one of which is
broken off, of an entirely different type. These long shoots
may be as much as 50 cm. long, and are provided with long
loriform leaves reaching a length of as much as 15 cm. I
feel almost certain that each of these long shoots is the
primary axis of a single plant, but as I have found two on
some of the specimens, chiefly in dry material, and have
not been able to satisfy myself that there were also two thalli
I must leave the question for future decision. The ana-
tomy is quite different from that of the floriferous shoots,
and very similar to that of the primary axis of Hydro-
bryum olivaceum.
This apparently primary axis, whose upper parts I only
know from the herbarium material, is non -floriferous as a
rule, though I have seen it with lateral tetrastichous flori-
feroLis branches. By the beginning of January, to judge from
the material collected by myself, it has usually died down,
and is only represented by a more or less decayed stump at
the base. This stump is often 5-8 mm. thick, and therefore
much stouter than the floriferous shoots. The primary axis
bears leaves in a complex phyllotaxy, which I have not been
able to make out. It is evidently very flexible, and drifts
out with the current like the shoots of Tristicha ramosissima
In transverse section it shows a central vascular strand.
f5U)
372 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
slightly asymmetrically placed as in Hydrobryum oliva-
ceum, and the anatomy of the tissues is also very similar to
that seen in the latter plant. There is a broad parenchy-
matous cortex, without any sheath of collenchyma round the
bundle. The vascular strand itself is made up of groups
of phloem-like tissue, divided by strands of long parenchy-
matous cells ; the groups are like those in the thallus of
Dicræa, and probably arise in the same way.
At the base of this axis is the thallus, which is com-
paratively very small ; it is crustaceous, like that presently
to be described in Hydrobryum, closely creeping, and
attached to the rock at all points. It is very difficult to make
out its mode of growth in the mature specimens, but it
appears to resemble that of Hydrobryum, and the thallus
is lobed like that of H. olivaceum. It bears (PI. XXX., fig.
1) numerous closely crowded secondary endogenous shoots,
which have already been mentioned as forming the most
striking character of this plant. These may be seen in all
stages, from very young ones just emerging from the thallus
up to fully formed flowering or fruiting stalks.
I have been unable to make out in what order or arrange-
ment the secondary shoots are developed upon the thallus.
The lobes of the thallus are very irregular, and very com-
monly there is a single line of secondary shoots along each,
but I am inclined to think that the original origin of the
shoots is rather more marginal than central.
Beddome, who first described this plant, figures it as
having two kinds of secondary shoots, the one such as we
have described, and the other purely vegetative, with
tetrastichous leaves without sheathing scales. Further in-
vestigation is very much required at early periods in the
life-history to determine the exact phenomena, and whether
all the secondary shoots are at first like the latter type.
Towards the end of the season most of the still submerged
secondary shoots bear scales on their whole length, but near
the top, as Beddome has also figured, these scales bear long
deciduous loriform green assimilating tips which drift out
OF CEYLON AND INDIA.
373
in the water. One is therefore tempted to suppose that the
scales arise late in the season, as in Dicræa, by the enlarge-
ment of the bases of ordinary leaves and the fall of their
long tips, and as some shoots are usually to be found (c/.
Beddome’s figure) with scales below and simple leaves
above, this is not altogether improbable. The quest! on must
be left for future settlement on the spot.
It is, however, certain that when a young secondary shoot
appears near the end of the season, as in PI. XXX., fig. 1, it
has scaly leaves from the first, and usually without any
deciduous tips. This phenomenon is not infrequent also in
Dicræa and Hydrobryum.
The anatomical features of the secondary axis are very
different from those of the primary. It stands stiffly erect,
at least in the flowering season, and this rigidity is given by
a stout belt of lignified tissue surrounding the central strand.
The structure of the latter, too, is quite different from that
of the strand in the primary axis, but this point must be
left for future description.
The scaly leaves fit together very closely, and the exposed
parts are hard and almost brittle v/ith silica. The long tips
always drop off on exposure to the air.
At the top of the secondary axis is the spathe (PI. XXX.,
fig. 3), usually of urceolate form, with two stiff teeth. It is
half buried among the uppermost leaves, and the spathe is
at right angles to the top pair of these. From this fact,
together with its two teeth, one is tempted to regard it as a
pair of leaves combined, but there is as yet no evidence of
this being the true explanation beyond that just given.
The exposed upper end is hard and stiff with silica, while
the included part is very thin and delicate. In my taxo-
nomic paper I have described the tip of the spathe as
circumscissile, but this is hardly quite the correct expression,
as there is no definite splitting layer of the spathe, but it
splits round so as to separate the siliceous upper part from
the non-siliceous lower. The upper part falls away as a cap
and exposes the flower.
374 WILLIS : MORPHOLOGY OP THE PODOSTEMACBÆ
The flower opens as soon as exposed to the air, and thus,
owing to the erectness of the secondary shoots, has generally
ripened its fruit by the time the whole plant is exposed
upon the dry rock. It is figured in PL XXX., fig. 4, and has
the usual structure of the flowers of this group of Podoste-
inaceæ, with stamens slightly exceeding the stigmas. It is
anemophilous, and apparently largely self-fertilized. It is
quite sessile among the upper leaves of the shoot. It ripens
quickly into a comparatively large smooth fruit, which has
a broad dehiscence rib only very faintly prominent on either
side, and a stouter rib in the centre of each valve (fig. 5).
One valve remains persistent on the stalk after dehiscence,
while the smaller and the seeds fall away. The persistent
valve often curves inwards as in fig. 6 in dry air. The wall
of the fruit (fig. 7) shows a very similar structure to that of
Griffithella, with stout sclerenchyma layers. The ripe fruit
has a smooth exterior, from which the parenchymatous
outer layers of the ovary wall have fallen away. It very
commonly stands on a long or short pedicel, owing to the
falling away of the upper leaves and the cortex of the stem.
The upper part of the pedicel is usually bifid, consisting of
the two vascular strands that formerly went to the ovary,
with their woody sheaths.
Rejuvenescence. — I have not observed any evidence of the
occurrence of this phenomenon, but in all probability it
occurs as usual.
Comparing Willisia with preceding genera, it is evident
that in some respects it is more on the level of Podostemon,
in that it has large and complex secondary shoots. On the
other hand, these bear each only one flower. The flowering
takes place while the water is still comparatively deep,
though the seeds are ultimately shed upon the rocks. In one
species of Podostemon we had an indication of a crustaceous
type of thallus, which, however, was more of the Dicræa
type ; in Willisia we get a crustaceous closely attached
OF CEYLON AND INDIA.
375
thallus, small it is true, but leading on to the type we shall
now have to consider in Hydrobrynm. The dorsiventrality
of this genus is not very marked, other than in the thallus ;
the flowers, however, are very dorsiventral in structure,
though the secondary axes stand erect ; as yet, however,
we only know the plant at its flowering season. It seems a
plant ill-adapted to water that is, or is liable to be, shallow.
HYDROBRYUM.
[(Endl.) Tul. ; Willis, Rev. Pod. Ind., Ann. Perad., I., p. 285.]
As explained in the preceding paper, I have defined this
genus, which has undergone many changes, practically in
the same sense as that of Tulasne. Its most interesting
species is perhaps H. olivaceum, which has caused much
confusion by its peculiar habit and mode of growth. It has
bf^en described by Warming (42, lY.) under its own name
from Ceylon material, and in the same paper under the name
Dicræa apicata from Nilgiri material ; neither description is
exhaustive, and there is much to add. I have been able to
study most of the species in the living condition, with the
exception of H. Johnsonii, of whose autonomy I feel doubtful,
and H. sessile, a new species discovered by Mr. Barber, but
of which he kindly sent me abundant spnrit material. The
genus is apparently confined to Asia, where it is common
in suitable places.
Hydrobryum lichenoides, Kurz.
(Plates XXVIII., XXXI., XXXII.)
This species occurs in numerous forms, nearly all of which
I have studied on living material, in Ceylon and India. The
plant shows a most remarkable amount of variation in
many of its characters, and these variations seem to be
different at almost every locality, so that it probably may be
divided into a great number of varieties, even if it be not
necessary to split it ultimately into several species. For the
376 WILLIS : MORPHOLOGY OP THE PODOSTEMACEÆ
purposes of the present paper, however, this variation does
not matter much, being chiefly in features that need only be
briefly touched upon,
Hahitat. — The plant affects smooth rocks in rapid currents,
but on the whole is not found in quite such violent water as
H. olivaceum. It occurs very commonly in quite small
streams, where the risk of shallow water is great, just as do
the other Hydrobryums and Lawia zeylanica, and perhaps
correlated with this is the fact that it occurs at much higher
€ilevations than most of the Indian Podostemaceæ. In Ceylon
it is rare, occurring only, so far as yet known, in one river,
at a height of 3,500-4,000 feet. In this river there is not, so
far as yet known, any other species of the family. In the
Sholai-aar, in the Anamalai mountains, it is commoner,
and is mixed with H. olivaceum, and also, as PI. XXVIII.
shows, with Willisia selaginoides. Mr. Barber found it in
Kanara, and I myself found it in enormous quantities every-
where in 1 he Khandala district, where it is most conspicuous,
and accompanies Lawia zeylanica in almost every stream.
It is difficult to understand how it has been overlooked in
this district, but in the Museum at Kew I found some
specimens from the Bombay Ghats collected by^ Law, and
labelled as Lawia pulchella. It is also frequent in the
Cherrapunji district at 4,500 feet, and often mixed with
Dicræa Wallichii, and I found it at Shillong and in the
foothills between Shillong and Gauhati.
Dry Seaso7i Apiwarance. — This is well shown in exposed
plants in PI. XXXI., and in plants from just below the
water-level in PI. XXYIII. Most of the plants in the first-
mentioned plate are from the Khandala district, where this
plant is to be seen in its greatest luxuriance and development.
The rock is covered with creeping thalli, as closely and firmly
attached as those of Lawia, a few millimetres broad, and
tapering towards the tip, brown, gray, blackish, or nearly
white, acccording to the form and the locality. They are
very regularly branched, alternately or sub-oppositely, with
little prosti'ate shoots in the forks of the thalli, from each of
OF CEYLON AND INDIA.
377
which arises one fruit, standing upon an erect axis emerging
from the boat-shaped spathe, which is split along the upper
side and usually lies nearly prostrate upon the rock. The
fruit is ribbed, and is usually open, with the lai’ger
valve persistent upon the stalk, while the smaller has fallen
off with the seeds. In specimens taken below the water, as
in PI. XXVIll., or in the upper right-hand figure of
PI. XXXI., the secondary shoots can be seen as little buds,
prostrate on the thallus ; those close to the tip are still often
in the leafy vegetative condition.
Germination and Life History. — I have only been able to
get two or three seedlings of this plant, at Hatton in Ceylon,
and all were at the age of the one figured in PI. XXXII., fig.
1. Evidently the germination takes place much as in Dicræa
stylosafucoides, giving rise to a short stout hypocotyledonary
axis with a few leaves at the top on a condensed epicotyle-
donary stem, the leaves arranged according to some
phyllotaxy which I have not been able to discover, but not
at any rate distichous. At the base of the seedling the thallus
emerges, and soon widens out at the tip and begins to branch.
The branching is more fully shown in fig. 2 ; at first it is
usually alternate, but later it is not infrequently sub-opposite,
as the photographs show. The growing apex is like that of
Dicræa, or like that of H. olivaceum described below
(excepting that it is an apex, and not a margin, as in the
latter), with a collenchymatous cap. The branching is close
to the apex, exogenous, and lateral as in Podostemon
subulatus, with a secondary shoot formed between the two
lobes in each case, at a very early period. Vascular bundles
are also developed leading to the branches of the thallus and
to the secondary shoots. The thallus grows rapidly in size,
and may reach a length of several inches during even the com-
paratively short vegetative period available in the Bombay
Ghats. The lateral branches themselves branch again, the
first branchlet being on the basiscopic side as usual. The
secondary shoots are like those of Dicræa, mere tufts of small
leaves emerging from little openings in the thallus.
378 WTLLTS : MORPHOLOGY OF THE PODOSTEMACEÆ
Later on, as the flowering season approaches, the develop-
ment of the flowei s takes place, most of the secondary shoots
producing each a single terminal flower. The axis elongates
and becomes prostrate on the thallus(or nearly erect in some
varieties), pointing in all cases towards the tip. The bracts,
2-8 in number, which form on each shoot, are developed
just like those of Dicræa, by the enlargement of the sheathing
bases of the leaves and the fall of the tips. The upper
exposed side of the sheath is thick and siliceous, the lower
thin and membranous. At the end of the shoot is the flower
enclosed in the usually prostrate spathe (PI. XXXIL, fig. 3).
When exposed to the air by the fall of the water (and thus,
owing to the very dwarf habit, almost only at a time when
the water is about to leave the plant altogether), the spathe
splits in a more or less irregular way on the upper side, and
the flower emerges on a short stalk and stands erect. Like
all the other flowers we have described, it seems anemophi-
lous and self-fertilized, with the chance of a cross at times
owing to the nearness of the flowers to one another on the
rock.
The structure of the flower has already been sufficiently
described. The chief point of importance to be noted here,
as bearing on the taxonomy, is the great variability to be
found among the stigmas, of which a few instances are figured
in PI. XXXIL, figs. 5, 6 (and cf. PL XXXVL, fig. 8). The
average form perhaps is ovate, but every stage may be found
from simple narrow subulate to broadly obcuneate with
many teeth. As the latter form of stigma is the character
on which Weddell chiefly bases his reduction of the genus
Hydrobryum to the solitary species H. Griffithii, and we
shall find the same variability in the stigmas of this species
also, it is evident that this character is not generic.
The fruit soon ripens, andis anisolobous with one deciduous
valve. It has eight ribs, well marked in most cases, but
almost evanescent in some of the varieties. In some, too, the
ribs are confluent some distance below the tip of the fruit,
in others only at the tip.
OF CEYLON AND INDIA.
379
Rejuvmescence."--^T]ii^ appears to be frequent in this
species and to occur in the usual way, by formation of new
thallus growing points.
Hydrobryum ssssile;, Willis.
'.Plate XXXII.)
This species was discovered in S. Kanara by Mr. Barber,
who kindly sent me a good supply of alcohol m.aterial. It is
of interest as showing many transition features to characters
which are well marked in other forms.
Habitat. — This appears to be very similar to that of the
preceding form — smooth rocks in rapid or shallow streams.
Dry Season Appearance. — This is very like that of H.
olivaceum, to be next described, but on closer examination
it is easily seen that the rock is not covered, as at first sight
appears to be the case, witha continuous coating of thallus, but
that there are narrow slits dividing the branches of the thallus.
The fruits are sessile among the dead bracts, and smooth.
Mature Structure. — In general construction and growth
the plant is like H. lichenoides. The thallus, however, is
very much broader, and the branches are so close together
that it practically covers the entire surface of the rock on
which it grows (PI. XXXII. , fig. 7). It thus forms a transition
in this feature to the true lichen-like thalli of the succeeding
species. The secondary shoots and the development of the
fioral buds are as usual. The flowers are sessile with long
stamens and curved ovary (figs. 8, 9). The fruit is smooth
and sessile, thus perhaps forming a transition to the fruit of
Farmeria metzgerioides described below. Rejuvenescence
probably takes place in the usual way.
Hydrcsbryym oSivaceiim (Oardii.)^ Tyi.
(Plates XXXII.-XXXVI.)
As already mentioned, this very remarkable plant has been
described under several different names, and I shall endea-
vour to explain below the origin of this confusion.
(51)
380 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
Habitat.— species is very abundant at Hakinda, where
most of my investigations have been made, and where it
covers large areas of rock to the exclusion of other forms,
though it is often found mixed with Lawia zeylanica, less
often with Farmeria metzgerioides or Dicræa stylosa
f ucoides, and only rarely with Podostemon subulatus, Dicræa
elongata, or D, stylosa laciniata. It affects on the whole the
most violent and rapid water of all the species, and is often
found on the edges of waterfalls and similar places of very
strong current. In the Nilgiris I found the var. griseum in
very similar places, and occasionally mixed with Dicræa
dichotoma, while in the Anamalais 1 found the local variety
mixed with Hydrobryum lichenoides, var. Fentonii. The
interminglingof this species with the larger plants mentioned
usually occurs mainly at places where there is considerable
local variation in the condition of the water, as for instance
at a place where an eddy rejoins the main rush of the stream.
This species, like the others with low-lying thalli, appears
to be able to inhabit much smaller streams and shallower
water than the large species, and is (consequently perhaps)
found at higher levels. It goes far higher than any other
species in Ceylon, except H. lichenoides, and is frequent
from 1,500 to 5,000 feet, and in quite small streams.
At Paikara in the Nilgiris it reaches nearly to 6,000 feet, and
in the Anamalais I found it at 3,500 feet, and probably
enough it may be found at higher elevations. In Ceylon it
occurs in very small streams, which easily run almost or
quite dry, but the thallus lies so low upon the rocks, and the
leaves of the secondary shoots form so good a sponge for
catching any water that may be trickling over the plant, that
it is able to live so long as there is any water, and it can
even stand a considerable period of complete exposure and
yet revive and form new growing points if once more
submerged.
Dry Season A2)2^earance,—T\xQ appearance of the plants
when completely exposed and in ripe fruit is very well
shown in PI. XXXIII., but the plants there figured are
OP CEYLON AND INDIA.
381
imnsually good examples of single plants, and show the
general form and lobing of the thallns in a way that is very
uncommon in most places, owing to the way in which the
thalli grow over one another, and to the frequent formation
of secondary growing points from broken or otherwise
injured parts. The rocks are covered with an irregular
crustaceous coating of dry easily-powdered thallus, gray to
white in colour, and with numerous short-stalked ribbed
fruits standing erect upon it, the stalks emerging from little
prostrate shoots of scale leaves on the thallus. Each pros-
trate shoot has a terminal boat-shaped spathe, splitting along
the upper margin, just as in the preceding species, and as can
be cl early seen in t he photograph. The thallus has an irregular
lobed mar gin, towards which the secondary shoots point; they
thus often, in very irregular portions, seem to point all
ways, and many of the artists of this plant have consequently
supposed the direction to be perfectly random, and have
produced very misleading pictures. The plant in the lower
left-hand corner of the plate shows clearly that the growth
of the thallus must have been from a common centre as in
Lawia.
If we examine at this period the plants which are still
submerged, we shall find that the thallus is green, olive, or
reddish in colour, with delicate margins, which are very
often more or less decayed or disorganized, especially in the
Indian forms. It is, as a rule, very hard to find specimens
showing any such regularity as that exhibited by those in
the figures, and this may account for the extreme inaccuracy
and misunderstanding of most descriptions of this plant.
The submerged thallus will be found to bear leaves, but these
of course soon fall when exposed, and hence the herbarium
specimens are leafiess and the plant is described as such,
even by those who have collected it alive. The leaves in
falling leave the scars of the secondary shoot upon the
thallus, as can be clearly seen in the photograph. Most of
the secondary shoots, however, have formed fioral buds by
this time and the leaf tips have fallen.
382 WILLIS : MORPHOLOGY OP THE PODOSTEMACEÆ
In the Nilgiri form, and occasionally in the others, the
simplicity of the construction is further masked by the
curious way in which the thallus frequently crumples, as
seen in the right-hand specimen in the plate, where the
upper parts of the thallus are crumpled and ridged into
a dense irregular mass.
The flowers open at once after exposure, and shed their
seeds upon the rocks in the usual way, where they may be
retained in crevices of the rock or of the old thallus.
Germination and Life History. — Germination takes place
in April and May, when the Avater-level rises during the
little monsoon. The first stage is much the same as in the
species already described. The seed SAvells and bursts its
coat, the hypocotyl emerging and bending downwards to
touch the substratum, where it promptly becomes fastened
by means of the usual rhizoids developed from its superficial
cells. The basal portion usually enlarges slightly, and from
it are almost immediately developed outgrowths — the thallus
and haptera. The early stages of development are extremely
interesting, and I much regret that I have not been able to
obtain sufflcient material to follow out all the details. The
seedlings are very minute, and grow in places where it is
very hard to And them, and I have to consider myself very
fortunate in having got so many as I have actually
obtained.
PI. XXXII., flg. 11, shows the earliest stage that I have
seen, the seedling being. about 2\ mm. high, with two
cotyledons developed, an erect hypocotyledonary axis as
loDg as the cotyledons, and two small outgrowths at the base.
In the specimen here shown one of these, the left-hand one,
was certainly exogenous, the other, on the right, endogenous.
It is often extremely difficult in these seedlings to make out
whether a particular organ is exo- or endo-genous. Those
cases where the organ is certainly exogenous are easy, for by
aid of the microscope the surface cells can be traced from
the stem to the lateral organ, cell fitting on to cell with no
break of continuity ; there are, however, many cases where
OF CEYLON AND INDIA.
383
this cannot satisfactorily be done, and sometimes it cannot
be decided what is the true state of affairs. Sometimes,
however, the endogeny can be made out by tracing the cells,
and at the junction of the two organs a few remains of cell-
walls or a break of continuity in size or mutual interweaving
of the cells may be seen. The endogeny, Avhen it occurs, is
usually only under one or two layers of cells, and as soon
as the organ emerges, its surface cells come to the same level
as those of the organ from which it springs, obliterating
almost completely the proofs of endogeny.
To return to the case in hand. The exogenous outgrowth
is a hapteron, the endogenous the first appearance of the
thallus. In about half of the few cases I have studied,
there was always one endogenous outgroAvth, which could
be recognized as thallus by the structure of its growing-
margin. In the remaining cases, the organ which cnild be
definitely stated to be thallus arose without doubt in an
exogenous manner. We have therefore in this plant a
transition case, the thallus sometimes forming exo-, some-
times endo-genously.
The seedling thus forms a thallus and one or more hap-
tera at the very earliest stage. Some of the figures show
other seedlings with more than one hapteron, but none of
the few cases that I was able to examine showed a formation
of more than one thallus, though I suspect that this does at
times occur.
The cotyledons are opposite to one another, with slightly
channelled upper sides, and are awl-shaped. They continue
to grow in size for a little while, as the various figures show.
The upper side is very hairy, with long unicellular hairs.
This is an interesting case of adaptation appearing in the
actual embryo, for the mature leaves of this, as of many other
species of this family, have hairy upper surfaces. Between
the cotyledons is visible a very minute plumule. The
figures above mentioned show several stages a little more
advanced. The seedlings shown are all, it will be noticed,
upon old capsules, for it was only in such positions, as a
384 WILLIS : MORPHOLOGY OP THE PODOSTEMACEÆ
rule, that I was able to find them under the water. All
show a thallus more or less developed, and one or more
haptera. Where the thallus has developed exogenously from
the hypocotyl, it is only to be distinguished from the hap-
tera, so long as it remains small, by the structure of its apex
or rather margin, which is on the whole broader, and more
evidently meristematic, the cells being smaller and more
numerous. The thallus nearly always starts in a downward
direction. It seems most frequently to arise at right angles
to the plane of the cotyledons, as the figures show.
By means of the haptera and the thallus, both of which
follow all irregularities of the substratum and develop
rhizoids on the lower side, the seedling is very firmly fasten-
ed to its support, and cannot be washed away by any force
of the water. The haptera seem usually to remain small,
whereas the thallus soon reaches a considerable size, and by
the end of the year may be a foot in diameter. The primary
axis is usually bent down by the force of the water, and the
cotyledons also often twist round to such an extent that the
hairy upper surfaces face downwards.
The first two leaves soon appear between the cotyledons,
in a plane at right angles to that of the latter, or nearly so
(fig. 13). They appear in alternate order, but very nearly
at the same time (fig. 12). The next two are approximately
at right angles to the first (fig. 13). The primary axis
continues to grow in length and thickness while this is
taking place. The leaves continue to appear at the apex for
some considerable time, and the axis to increase in size to
correspond (PI. XXXIY., fig. 3), till at last, towards the end of
September, it may be as much as 5 cm. high and 3 mm. thick
with a large number of leaves at its apex, which themselves
may be 10 cm. long (Pi. XXXIY., fig. 4). When a number of
leaves have been formed, the phyllotaxy seems to be of the
I or type, but I have not been able to make this out in a
satisfactory way from want of material, and I do not think
the point of very great importance. PL XXXII., figs. 15, 16,
shows a young plant with its leaf arrangement. In a short
OF CEYLON AND INDIA.
time the original growing point at the apex of the stem
becomes replaced by three or more, as in figure 17, which
has three, and as time goes on there may be a large number of
growing points present. The branches thus formed, how-
ever, never elongate, and the stem always remains as a simple
axis, crowned by a tuft of leaves.
The first formed leaves are very small, not more as a rule
than 5-10 mm. long. As the stem grows the successive new
leaves are larger and larger, till the last formed ones may
reach 10 cm. in length and 1 mm. in diameter. The leaves
are hairy on the upper side, as in other species of this genus,
and exactly resemble those borne on the endogenous shoots
of the thallus, to be described below. They are perfectly
simple, long, and very narrow, flat or slightly hollow on the
upper side, convex upon the lower, needle-shaped with very
acute apex, and slightly sheathing at the base.
In transverse section the structure agrees with that
described by Warming (Dicræa apicata, 42, IV., p. 155, and
fig. 24). The leaf has a small vascular bundle in the centre,
apparently consisting only of phloem tissue.
It is worthy of note that though the primary axis of
the plant in this species thus becomes of considerable
importance, more so than in most of the other species that
have been studied, it still consists, throughout its life,
practically only of the hypocotyl. The appearance of the
primary axis when fully grown is shown by PI. XXXIV.,
fig. 4, which is a modified reproduction of the original
drawing of Podostemon Gardner!, Harvey.
These leafy primary shoots are the most important
assimilatory organs of the plant up to the end of August or
September, when the thallus begins to be very large (PI.
XXXV. shows an entire plant at this stage). About this
time they usually begin to suffer from the water pressure,
and from the accumulation of rubbish which usually clings
to them, and they often get broken off or flattened down
upon the surface of the rock or the thallus, and the leaves
tend to break off. When bent down they frequently become
3(S() WILLIS : MORPHOLOGY OF THE PODOSTBMACEÆ
obliterated by a piece of thallus growing over them. They
may still be commonly seen at the end of October, and
traces may be found even in January, so that we may say
that they survive through the whole life of the plant.
The internal anatomy of the stem can only be briefly
touched upon ; it has already been described by Warming
(Dicræa apicata, 1. c,). Towards the middle of the cross
section, but a little excentrically, is a vascular bundle.
This is not very clearly marked ofl: from the surrounding
tissues, and shows a somewhat simple structure, like that
of the primary axis of Willisia, consisting when mature of
a number of irregularly arranged groups of phloem tissue
with thicker walled cells between the groups. Details must
be left for subsequent description.
To return to the thallus. It grows outwards in close
contact with the rock, to which it becomes attached by the
usual rhizoids, or in the Nilgiri form and occasionally in
the others by haptera also. Only the extreme marginal
portion, for about 2 mm. or perhaps at times 5 mm. from the
edge, is actually growing and expanding, so that the attach-
ment can soon be made without risk of severance by further
growth. In a very short time the thallus shows a scallop-
shell form (PI. XXXII , fig. 14). When cut in longitudinal
section (PI. XXXIV., fig. 2), the edge of the thallus shows a
meristem like that of Dicræa, with a collenchy matous root-
cap at the outer part, which is derived from a more or less
regular transverse meristem that on the inner side gives rise
to the thallus tissues. The root-cap is usually slightly
raised above the rock, and hardly seems to perform any
protective function in preserving the meristem from contact
with the substratum, though it must be of a certain
protective value when the thallus collides with another
thallus or with a projecting portion of rock or other obstacle.
It is possible also that it may have a stiffening function like
the rim of bundles along the margin of a dicotyledonous
leaf. The presence of this tissue is almost the only constant
“ root ” character, whether physiological or morphological.
OP CEYLON AND INDIA.
387
that remains to the thallus of this species, though a series
may be easily traced back to the root-thalli of such forms as
Tristicha. As seen in transverse section, the thallus shows
very similar features to those of most of the other genera
with flat thalli that we have already examined. It has (PI.
XXXIY., fig. 5) a well-marked epidermis and parenchy-
matous cortical tissues, in the lower part of which the
vascular bundles are to be found, showing the usual structure,
but very small. The course of the vascular bundles in the
thallus and their relation to the groAving margins is very
complex, and I have not been able to make out satisfactorily
any general principle. Each secondary shoot of course has
a vascular bundle to it.
The thallus does not grow in thickness like that of a
Dicræa or a Grifhthella, though it becomes a little stouter
behind the margin by the increase in size of the cortical
cells. It is very sensitive to contact or to light or gra\dty,
and folloAVS out every irregularity of the substratum.
Before its diameter has reached more than a few milli-
metres the thallus begins to branch ; this process seems to
take place by a slowing of the growth in certain spots, and
the thallus becomes lobed. When growing on smooth rock
the thallus usually assumes a sort of trefoil shape in a short
time, but very commonly there is no recognizable symmetry
about it, as it grows to fit the more or less irregular sub-
stratum, and usually comes into contact with other plants.
Each lobe as formed continues to groAv in the same manner,
and so does the tissue at the base of the fork between the
lobes, so that, though these forks are at first quite close to the
primary axis, later in the year there are none within several
inches of it. The primary axis at first is at one side of the
plant, but very often the thallus overlaps round it, or in many
cases the tissue at the base of the stem grows out like the
thallus, so that the stem stands more or less in the centre of
a thallus. The lobes as they grow in size go on branching in
the same way as at first, so that at last the appearance of a
plant growing in an isolated position upon a large piece of
‘ (52)
388 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
smooth rock (PI. XXXV.) is like that of a large more or less
circular piece of lichen, lobecl along the margin, the lobes
being rarely more than an inch deep, and the division
between them almost obliterated by the fact that the thallus in
its growth does not diverge at the bases of the divisions, but
if anything tends to converge, so that the lobes oftener than
not overlap each other. The diameter of the thallus reaches
15-18 cm. (6-7 in. 1 by the end of August (PI. XXXV.), and
by January, when its growth finally ceases, often as much
as 30-36 cm. (12-15 in.). The growing edge of the thallus is
usually of a more or less deep red colour, while the mature
part is perhaps most often of an olive green, but in dry
weather, when the water falls so as to expose it to more
intense light, it very often assumes, like all the other species,
a reddish colour all over, due to the presence of anthocyan
in the epidermal cells. As a rule, these plants have to grow
on very irregular surfaces of rock, and so the thallus becomes
equally irregular in shape. Often, too, a number of seed-
lings commence close together upon a rock, and as they
grow they collide with one another, and one grows over
another ; when this is the case the lower thallus soon dies.
Insect larvæ of various sorts feed largely upon the thalli,
which contain enormous quantities of starch, and thus help
to destroy their symmetry. If the water falls, as is not in-
frequently the case in the drier end of the south-west mon-
soon, in August and September, so as to expose the thallus,
the exposed parts soon die, especially the growing margins,
and when this is the case, the symmetry may be almost
wholly lost; the growth is recommenced when the water rises
again, in any part of the thallus that has not been killed, by
a process of rejuvenescence described below, but the former
symmetry is so far gone that little remains to show the
principle of growth of the thallus. Hence, when the water
finally exposes the plant in January, the principle of growth
cannot be clearly made out at all, except by good fortune in
finding plants that have not suffered any of the accidents
we have described, and it is not surprising therefore that
OF CEYLON AND INDIA.
389
under the circumstances it should be described as having no
definite form.
Not only Joes this plant bear a large number of big leaves
on its primary axis, but it also produces a great number of
leafy secondary shoots upon the thallus, so that it could
hardly be more incorrectly described than has hitherto been
done Before it has reached a diameter of more than one
centimetre, it begins to produce these leafy shoots, two of
which are shown in PL XXXIY., fig. 1, upon a very young
seedling. The seedling shown in the figure is clasping an
old capsule, and has already become lobed. The primary
axis had become broken olf very early, and this no doubt
accounts for the large size of these leafy shoots upon so
small a thallus, for, as a rule, they have hardly begun to make
their appearance above the surface in a thallus of this size
{cf. fig. 3). Their first appearance is seen in PI. XXXVI.,
fig. 3 {an.). They appear in the zone of growth close to the
margin of the thallus, as little round markings on the upper
surface. A. trifling distance further back they pass into the
region of elongation of cells, and become elliptical. The
centre of the mark is usually green, but the cells round it
as a rule show a deep red colour, like those of the margin
of the thallus. Soon afterwards the first leaf of the shoot
breaks through the superficial cells and appears above the
surface of the thallus, so that in this species, as in all the
others, the leafy shoots borne upon the roots are' endo-
genous. A second leaf soon follows the first, and the plane
of the two leaves is at right angles to the margin of the
thallus. More leaves soon appear, till the shoot, when fully
grown (PI. XXXIV., fig. 4), may consist of leaves as much as
5 cm. long, and five, six, or more in number. These leaves
exhibit the same structure and appearance, though usually
on a smaller scale, as the leaves of the primary shoot. The
growing point of the shoot does not at first elongate or
appear above the surface of the thallus, and there is, perhaps
in consequence of this fact, no appreciable dorsiventrality in
the shoots at this period of the life-history. They remain
390 WILLIS : MORPHOLOGY OF THE PODOSTEMACBÆ
in this condition till a little while before the beginning of
the north-east monsoon in October, when they mostly
become trausformed into flowering shoots during the rainy
weather which lasts through the remaining part of the year,
till when the water falls in January only those near the
margin of the thallus still retain their leafy condition. The
flowering shoot consists only, as a rule, of a terminal flower
with a few bracts below it, and lies horizontally upon the
thallus, thus showing a dorsiventrality, which is absent in
the younger stages. The growing points, too, are then above
the surface of the thallus. PI. XXXIY., fig. 6, shows stages
in the transformation of leafy to floral shoots, which is very
similar to that already described in Dicræa. The floral shoot,
as was mentioned in describing the dry season appearance
of the plants, points towards the edge of the thallus. The
sheathing bases of the leaves now produced enlarge and
become fleshy, and presently the tips wither and fall off,
leaving the swollen basal parts of the leaves, which then
form the scales or bracts of the floral shoot. These scales
show a distinct dorsiventrality, in that they are very much
thicker in the exposed portions upon the upper than upon
the lower side ; the latter is quite thin and membranous,
while the former is thick and fleshy. This difference is
much less in species like Dicræa elongata, where, though the
shoots have an upper and a lower side, both are freely
exposed in the water.
At the tip of the floral shoot the flower develops, enclos-
ed in its spathe, and it is fully formed in all its parts early
in the north-east monsoon, though, perhaps for want of the
necessary food supply from the thallus, it seems incapable
of further development and anthesis in the event of the
water-level falling prematurely before December.
Should the fall of the water be somewhat early, specimens
may frequently be found in the condition shown in PI.
XXXIY., fig. 6, with the long green tips still persistent upon
some of the bract-scales, and in the specimens of Podostemon
Gardneri preserved in the herbaria this state of things is
OF CEYLON AND INDIA.
391
frequent (see also Prof. Warming’s figures and descriptions
of Dicræa apicata, which is really the species we are
now considering, as is elsewhere explained). As a rule,
however, before January, the season when the plants are
normally exposed to the air, all these tips have fallen, and
the buds present the appearance described in systematic
works, and shown in PI. XXXYI., fig. 1, which represents one
bud with the flower just exposed by the longitudinal splitting
of the upper surface of the spathe. At this period also,
as has been already mentioned, the primary axes are mostly
more or less complet- ly disorganized, and the thalli them-
selves have lost their simplicity of form to a very large
extent, so that when the final stage in the life-history is
reached, the plant presents very few salient features giving
any clue to the somewhat complex development through
which it has gone, and a description of it from specimens
collected at this time is almost certain to err in many points,
and to give no idea of the real appearance of the plant when
actually in the course of its growth.
The spathes split on contact with the air, and in a few
hours the flowers have emerged and are standing erect on
short stalks, ready for fertilization (PI. XXXVI., fig. 2). Like
those of the preceding forms, they are wind-fertilized, and
apparently largely autogamous ; the pollen is loose and
powdery, and may be seen blowing out of the anthers in
every gust of wind.
The fruit ripens rapidly, and in about a week after fertili-
zation the seeds may be shed. The thallus appears to be
able to live exposed to the air for a considerable time, but
ultimately dies if not again submerged.
Before leaving the consideration of this plant, a brief
description must be given of the very curious monstrosity
figured in PI. XXXYI., fig. 5. This was found on 20th July,
1898, among some seedlings. The primary axis has been
bent down, to show the rest clearly. The thallus is rather
thick, folded up into a kind of Y shape, and with a large
number of growing points round the margin, which in the
392 WILLIS : MORPHOLOGY OF THE PODOSTEMxlCBÆ
two cases marked g have grown outwards to some dis-
tance. From near one end there springs the trumpet-shaped
organ which is flattened out at the end, and appears to
have been attached to the rock (the specimen was obtained
in the usual way by groping in deep water) ; its margin
is like that of a normal thallus and appears to have grown
in a similar way. On the far side are four pitcher-like
organs, a, h, c, springing from the margin, in different
stages of development, and on them endogenous shoots
(s.s.) are forming, somewhat as in Dicræa elongata. These
pitchers are hollow at the outer ends for some depth. The
whole is of interest as showing once again the almost un-
limited plasticity of the thallus in these plants.
Rejuvenescence, — As in most of the other species, so here,
the thallus, so long as alive, seems to have an almost unlimited
capacity for the production of new growing points and the
renewal of growth after exposure to air or other injury.
The new growing point is formed endogenously behind the
injured portion, and grows out from it in the usual scallop-
shell form (PI. XXXVI., figs. 3, 4). Exactly how long the
thallus can stand exposure without losing all capacity for
regeneration is uncertain, but 1 have found several cases in
this as in other species, where it appeared to have survived
several months after being exposed by the fall of the water
in .January. When the flowers are ready to open the thallus
and stalks are crammed with reserves, especially starch, and
hence the ripening of the fruit is very rapidly carried out
after fertilization ; but all the starch is not thus used up,
and by means of Avhat is lefc the plant is able to regenerate
its thallus if placed under water once more.
With regard to Rodostemon Gardneri^ Harv. MS., a few
notes must be made. Thwaites’ original description is as
follows : “ caule simplici, terete, glabro,foliis plurimis capil-
laceis coronato. C. P. 2,989. Hab. on the membranaceous
rhizomes of Hydrobryum olivaceum. Tub In a rapid
OF CEYLON AND INDIA.
:^93
mountain stream at Ramboda, October, 1853, Dr, W. H.
Harvey.” He suggests that it may be an early stage of growth
of H. olivaceum. No doubt can remain that it is simply the
primary axis of the form just described, but to make
quite sure I have myself examined the spot where it was
found, and verified that the stream is inhabited only by
Hydrobryum olivaceum. I have also examined the her-
barium material at Peradeniya, Kew, and Paris, and in each
case found the thallus attached to the base of the primary
axis, leaving no possible doubt as to identity.
With regard to the Anamalai form, it need only be
remarked that it agrees in all essential respects with the
description of the Ceylon form given above, but the lobing
of the thallus is usually much deeper, more like that of the
Nilgiri variety.
The Nilgiri form, Podostemon griseus of Gardner, shows
considerable differences from the Ceylon form. The pri-
mary axis is larger and stouter, and sometimes bears flowers
either on the hypocotyledonary portion, or among the leaves.
Owing to the decayed condition of my material, I am
uncertain whether these flowers are borne directly upon the
axis, or, as in one or two cases seems to be the truth, upon
very small thalli formed upon the axis. The question should
be settled on living material in the latter part of the
vegetative season. The tli alius itself has a curious gray
colour in January, instead of the olive green of the Ceylon
form, and at this period it is usually found to be more or
less decayed and worn away, except just round the bases of
the shoots. It is, as the specimens in PI. XXXIII. show,
much more deeply lobed, as a rule, than the Ceylon form.
Its primary axis was described by Tulasne as Dicræa apicata,
whose identity I have verified as with Podostemon Gardneri.
Hydrobryum «lohnsonii (Wight), Willis.
As mentioned in the preceding paper, the autonomy of
this species is very doubtful, I am inclined to think that
it will ultimately prove to be the same as that already
»^94 WILLIS : MOLPHOLOGY OF THE PODOSTEMACEÆ
described as H. lichenoides, Kurz, in which case of course
the latter name must become a synonym. In view of this
uncertainty there is no need to deal further with it here.
Hydrobryum GriffHhii (Wall. MS.), Tul.
(Plate XXXVI., figs. 6-8.)
This species is the only representative of Tulasne’s sub-
genus Euhydrobryum, subsequently raised to generic rank
by Weddell, and characterized from its twelve-ribbed
isolobous fruit and fan-like stigmas.
Habitat — I found it growing on rocks on the bank of the
Kalapani in the Khasia mountains, over which a slight
trickle of water was running, but which from their extreme
steepness would probably have a very rapid current in the
rains, though perhaps not a very deep one. The habitat was
almost exactly similar to those in which I have often found
H. lichenoides and H. olivaceum.
Dry Season Appearance. — This is very similar to that of
H. olivaceum, but the thallus is in general smaller, the buds
or shoots upon it are arranged in a more regular way,
evidently radiating from a common centre, and the isolobous
fruits are nearly prostrate, instead of being, as one might
expect from their symmetry, erect.
Mature Structure. — This is almost exactly like that of
H. olivaceum in all important points. The thallus is similar
in form to that of the Nilgiri variety of the latter, with deep
sinuses between the lobes, and with the same curious
growing margin. The secondary shoots are similar to those
of H. olivaceum, and the flower emerges in the same way.
The spathe shows a tendency to be bifid at the tip like that
of Willisia, and the exposed part of it is very siliceous, as in
that species, and seems sometimes to break off in a similar
way. The structure of the flower shows little of special
interest, exceptingin the stigmas, which exhibitmuch variety,
parallelling that of H. lichenoides, and completely destroying
the generic value of the stigmatic characters. In FI. XXXVI.,
OP CEYLON AND INDIA.
395
fig. 8, a few examples are shown, selected almost at random
from the few flowers at my disposal, and showing all forms
from simple subulate to the obcuneate and deltoid toothed
forms like those of the American Lophogyne.
The fruit is almost, but not quite, isolobous, and has
usually twelve, but often more, ribs. The extra four ribs, as
compared Avith the fruits of the other species, are intercalated,
and often do not run up to the top of the fruit, much as is
the case in the extra ribs often found in the Burmese form
of Dicræa Wallichii. This difference alone is certainly not
sufficient toalloAv of this species being generically separated
from the other Hydrobryums, with Avhich it agrees so closely
in thallus morphology, spathe, and other floral characters.
The genus Hydrobryum thus contains some of the most
peculiar of the many peculiar plants that we have described.
If only the species last described, H. olivaceum and H. Griffi-
thii, were knoAvn, we might be somewhat puzzled to know^
what to make of them, but the intermediate steps afforded
by H. sessile and H. lichenoides lead back Avithout any serious
discontinuity to thethalli in the forms already dealt wdth.
The dwarfing of the secondary shoots, and the enlargement
of the thallus to do the work of assimilation or to produce
more secondary shoots for that purpose, which we have
already seen in Dicræa, &c., is carried to an extreme in this
genus, and Ave get here forms as dAvarf in every respect as
Lawia itself. Probably directly correlated Avith this is the
fact that the tAvo genera have very similar habitats, and that
the species are very often mixed with one another. Both
inhabit water that is liable to become shallow very quickly,
and consequently perhaps are the genera which are found
at the greatest elevations, i.e.^ on the whole in the smallest
streams, and in the most northern localities, i.e., in the
rivers which most rapidly run dry. During the early part
of the vegetative season the Avater is usually deep enough.
396 WILLTS : MORPHOLOGY OF THE PODOSTEMACEÆ
and we find in H. olivacennij and perhaps in others, a com-
paratively large primary axis at this period. By the time the
risk of shallow water becomes great, the thallns is fully
established and the plant can do without the large axis. We
cannot regard the thallus as merely an adaptation for attach-
ment to the rock, for the primary axis is able to hold on in
the swiftest current if all the thallus be removed, except the
part immediately at its base.
In H. lichenoides we see the stage that would be reached
by a simple dwarfing of the secondary shoots of Podostemon
subulatus or better of P. Barberi. In H. sessile we get a
further stage in the broadening of the thallus, so that it covers
practically the entire surface of the rock included in its
outer outline. In this species the sinuses at times tend to be
obliterated by growth of their bases, and if this process be
carried a stage further, we get the deeply lobed thallus of
H. Griffithii and H. olivaceum griseum, and finally the
shallowly lobed H. olivaceum zeylanicum, in which growth
is no longer apical but marginal, and the root-cap is conti-
nuous round the whole outer edge of the plant. This thallus
shares with that of Dicræa the distinction of being probably
the most remarkable organ yet described under the general
morphological category of “ roots,” and were there not the
series of stages connecting it to that of Tristicha and
Podostemon, it might easily be looked upon as an organ of
an entirely peculiar class. We shall return to this subject in
the general summing up.
FARMEBIA.
[Willis, Rev. Pod. Ind., Ann. Perad., I., p. 246.]
This genus, though evidently closely allied to Hydro-
bryum, and resembling that in many points of structure, yet
differs in certain very marked peculiarities, and is of
particular interest. It is confined, so far as yet known, to
Ceylon and the extreme south of India, in each of which
there is a single species. The Ceylon species was described
OF CEYLON AND INDIA.
397
by Dr. Trimen, but incompletely, as he overlooked the
presence of the fruit in his material ; it is buried among the
bracts, and is indéhiscent, two almost unique features in this
order. Large quantities of fruit may be found in the material
distributed by him. Of the Indian species, I have only very
incomplete material at my disposal.
Farmeria metzgerioides (Trimen), Willis.
(Plates XXXVI.-XXXVIII.)
Hahitat. — This plant is extremely abundant at Hakinda,
where it grows on smooth rocks in eddies and rapids, as a
rule not affecting such swiftly moving water as does Hydro-
bryum olivaceum. As usual, it is most often found alone,
but it is frequently mixed with Lawia zey lanica, Hydrobry urn
olivaceum, and rarely with Podostemon subulatus or others.
It also grows in more shady places than the other species;
I have found it in quantity along the shady right bank of the
river below the corner at Hakinda, where the water is
overhung by Pandani, Fici, &c., and where the sun only
shines for a few hours in the morning.
Dry Season Appearance. — This is shown in PI. XXXVIII.
on the two stones in the middle. There is a narrow ribbon-
like thallus, upcurved at the edges, like that of Hydrobryum
lichenoides in most respects, branched right and left, but
more sparingly and not so close to the apex, and with
secondary shoots along the edges, these consisting of a few
little bracts like those of Hydrobryum, and with the fruit
concealed among them. Where a branching occurs, the
secondary shoot is behind, not in the axil of, the branch.
Germination and Life History. — When the water-level
falls at the beginning of the dry season, the plant may be
found in flower, e.^., I noticed it in flower at Hakinda
in 9th January, 1^98, 8th January, 1899, and 17th December,
1899. The flowers are sessile, and ripen into sessile fruits,
each containing two seeds only, which are larger than those
of most of the order. The fruit does not dehisce, but
398 WILLIS : MORPHOLOGY OF THE PODOSTFMACEÆ
remains tightly held down within the hard persistent bracts,
and the seeds germinate in situ and break through the thin
membranous fruit wall. The embryo (PI. XXXVI., fig. 9) is
short, straight, and thick, and from 0*5 to 0*75 mm. long, of
which length the cotjdedons form two-thirds.
The germination is fairly rapid. The hypocotyl emerges
from the testa and bends downwards to the substratum,
becoming attached in the usual way, w^hile its base becomes
somewhat flattened or tuberous (PI. XXXYI., fig. 10).
Haptera are rare ; one case was found with haptera at the base
of the foot formed by the hypocotyl. Fig. 14 shows two
seedlings emerging from a capsule more or less embedded
in the remains of the old thallus, bracts, &c.
The cotyledons spread out at once ; they are not opposite
to one another in this species, but stand at an angle of about
135°, so soon as they are open. Thus the dorsiventrality of
the mature plant shows itself in the very earliest stages of its
life. The notch between the cotyledon-bases runs lower-
down, i.e.y nearer to the foot of the hypocotyl, on the anterior
than on the posterior side of the expanded embryo, but I am
not able to make out whether the same is the case in the
ripe embryo in the seed. The wider angle between the
cotyledons is usually upwards, the backs of the cotyledons
being almost appressed to the substratum. The cotyledons
are wider and stouter in proportion than ordinary leaves,
and are not hairy. The apex is acute, and the base vertically
thickened as may be seen in the figures ; the upper side is
very slightly concave.
Figure 10 shows a very young seedling ; the cotyledons are
spread out with their lower sides almost against the ground,
and in the larger angle between them the thallus is emerging,
slightly above the base of the flattened “ foot.” The
direction of starting does not, as one might expect, bisect
the large angfe between the cotyledons, but lies as a rule in
a line with and opposite to one of the cotyledons, as may be
seen in several of the figures, especially well in fig. 11. In
some cases two thalli are formed from one hypocotyl, as
OF CEYLON AND TNDIA.
399
seen in figures 13, 15, 13. Tlie second thallus is usually in
the smaller angle between the cotyledons. In one case three
thalli were observed. The development of the thallus is
endogenous, but as usual this was a matter of difficulty to
observe, and in one or two examples it was impossible to
make definitely sure. The endogeny seems to be only under
one or two layers of cells. As soon as the thallus emerges the
line of demarcation becomes very indistinct, the cells at the
edge of the thallus coming into the same level as those
of the hypocotyl.
Leaves soon develop above the cotyledons. The first
usually appears in the larger angle between them, and the
second in the smaller, but apparently the re verse is sometimes
the case. Both cases are seen in the two seedlings of
figure 14. Subsequent leaves are arranged in an approxi-
mately distichous way upon the stem, the third leaf being
approximately over the first. The leaves are subulate, and
have sheathing bases, and bear hairs upon their upper
surfaces, like so many in this family. The growth of the
primary stem is not long continued ; it rarely forms more
than 6-8 leaves, of about 5-7 mm. in length. This primary
axis can hardly therefore be of very much use to the plant
after the thallus has become of appreciable size, and indeed
the primary axes, so far as could be found, seem to disappear
after the first two or three months.
To return to the growth of the thallus. When it first
emerges from the hypocotyl, it has a narrow apex, but it
soon widens into an ovate form, as in figure 11, and this form
of the apex remains essentially unchanged throughout the
life of the plant.
Examination of the apex of the thallus shows that the
extreme tip bears, or rather is composed of, thick-walled
collenchy matous cells, much as in Hydrobryum olivaceum,
much larger than the small and thin-walled and densely
protoplasmic cells immediately behind the tip. The latter
are meristematic, and the collenchymatous tissue represents
a root-cap, though it does not seem to be of much practical
400 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
value to the thallus in that respect. The cells of this tissue,
as they approach the edge, become longitudinally stretched
(/.6., tangentially to the surface of the growing point), and
now and then thin layers of them may be seen splitting off
from the end of the thallus, just like an ordinary root-cap.
The growth of the thallus is fairly rapid. As soon as it has
grown a few millimetres out from the hypocotyl, it becomes
attached by rhizoids to the substratum, and the endogenous
leafy shoots begin to appear. PI. XXXYI., fig. 12, shows this
stage : on the left side of the thallus, near the edge, a leafy
shoot is just emerging, while a little nearer to tlie apex on
the right-hand side is the swelling indicating the formation
of a second younger one. The shoots appear with great
regularity in this alternate and acropetal succession, when
not interfered with by any obstacles, but when the substratum
is very irregular, or there are many plants competing for
space, the regularity may be lost. The next stage is exhibited
in PL XXXVII., fig. 1. The primary axis shows the coty-
ledons and four leaves, while the thallus, springing in a line
with one of the cotyledons, has grown out to a considerable
distance from the hypocotyl, and has developed several leafy
shoots. At each node it is wider and also somewhat thicker
than at the intervening part. This specimen also shows the
branching of the thallus itself. The leafy shoots appear
first, and are often, but not always, followed at the nodes by
branching of the thallus. The shoots appear under two or
three layers of cells at a very early stage. The appearance
of the branches of the root-thallus is later, and it is often hard
to determine whether they are really endogenous. Their
development in the meristem remains to be Avorked out.
They appear on the under side of the thallus, near to but not
quite at the edge, just in front of the leafy shoot at the same
node, as seen in PI. XXXVII., fig. 3. They are endogenous
under a fe\v layers of cells. The branch thallus repeats the
structure of the main thallus, and does not as a rule stop
short in its growth like that of Podostemon subulatus, though
it often becomes stopped by obstacles.
OF CEYLON AND INDIA.
401
The thallus in cross section is convex above, and shows a
general construction like that of Hydrobryum lichenoides or
H. olivacenm. The leafy shoot developes a few small leaves,
rarely more than five or six, of about 5 to 10 mm. in length,
and hairy above.
The growth of the seedling plants is fairly rapid. For
example, on 21st May, 1898, the seeds were just germinating,
and on 19th July the rocks were thickly carpeted with plants,
many of which were from two to three inches long and
much branched. During the next four or five months the
plants seem to be purely vegetative. The flowers, when
they begin to form, develop very rapidly. The usual period
for this process appears to be December. Nearly all of the
leafy shoots then form flowers. By this time the plants have
grown to great lengths, and form a dense interwoven felt-
work upon the rocks, as may be seen in hitherto existing
herbarium specimens, Avhich as distributed consist of
fragments of very many different plants inextricably mixed.
PI. XXXVII., fig. 2, shows a thallus, which is just forming
flowering shoots in the middle of December. The apex is
still growing, and forming leafy sheets as usual, while a
little further back the flowering shoots occur in various
stages of development. At the tip of a root the flower shoot
apparently develops as figured here, without any long period
during which the axis is purely vegetative, while further
back on the thallus the old vegetative shoots may be seen
changing to flowering shoots. In either case, however, the
process is essentially the same. While the shoot is only
leafy, the growing point lies below the surface of the thallus,
but now it begins to elongate and come above it. The
leaves, hitherto thin and comparatively simple, grow out
at their bases, changing function and form like those of
a lily bulb ; the sheathing part of the leaf thus becomes much
enlarged. The axis remains in the horizontal position, or
rather the position parallel to the substratum, which in this
species it occupies from the very beginning. The upper side
of each sheath becomes fleshy, whereas the lower retoains
402 WILLIS: MORPHOLOGY OP THE PODOSTBMACEÆ
thin and membranons. Bracts are thus formed as in Hydro-
bryum. In that genus the free blade of the leaf usually
falls off early, so that it can rarely be found in January, but
in Farmeria this is not, as a rule, the case, and the bracts
have leafy ends until the water falls so low as to expose them
to the air, when the free ends wither and fall off. In the
centre of the bud thus formed by the bracts the flower
gradually develops within its spathe. By the end of the
wet season in December or January it is ready to open as
soon as exposed by the fall of the water.
The flower is quite prostrate, as in Hydrobryum sessile,
only the stigmas and the long stamen emerging and bending
upwards out of the spathe, which splits like that of Hydro-
bryum on the upper side. The stigmas are very long and
subulate. The stamen is single, as is occasionally the case in
Hydrobryum (PI. XXXYlI.,fig. 4). The chief interest of the
flower is in the ovary, which is quite different from that of
the rest of the order, so far as yet known. It is bilocular,
but much more dorsiventral than any other, the lower
loculus being more or less abortive, without ovules, and
displaced towards the stylar end in the ovary (PI. XXXYII.,
hg. 9). From the base of the ovary springs a short stalk,
which bears a large swollen placenta tilling the distal end
of the upper loculus of the ovary, the septa being on its
lower side and converging towards the stalk. The basal end
of the placenta is flattened (figs. 5, 6), and bears two large
ovules filling up the basal end of the loculus, themselves
flattened against the placenta (fig. 7). The tissues of the
placenta are closely packed with starch, and there is thus an
abundant supply of food ready for the ripening of the seed,
as the bulk of the placenta is quite as great as that of the two
ovules or the seeds which result from them.
The flower is as usual anemophilous. The ovary ripens to
a fruit of about the same size as itself (fig. 8) with two large
seeds, borne on the rather shrivelled placenta. The pericarp
is thin and membranous, and does not dehisce. As the fruit
OF CEYLON AND INDJA.
403
is quite enclosed in the persistent hard scaly bracts, the
seeds are thus firmly held against the rock, and ultimately
germinate in that position, as we have already seen. This
species is consequently one in which there is no great
difficulty in obtaining numbers of seedlings if the water-
level be only sufficiently low.
Farmeria metzgerioides is thus, so far as we as yet know,
the solitary case among the Podostemaceæ of the possession
of an adaptation to ensure the anchorage of the seeds in a
position suitable for their germination and growth. The
species is, perhaps in consequence of this, extremely abun-
dant at Hakinda, perhaps the most abundant of all. On the
other hand, the chance of distribution to a distance is
probably very much diminished. However, Farmerias in
vegetative condition, which appear to be specifically identical
with the Hakinda form, have been found at other localities
in Ceylon, some at a great distance, so that, unless we
suppose the species to have evolved separately at each
locality, we must assume that distribution to a distance is
at times effected.
Farmeria iiidica, Willis.
(Plates XXXVII., XXXVIII.)
Of this species, I have only material collected in the dry
season, on exposed rocks, and can only mention the most
interesting points of difference from F. metzgerioides. As
PI. XXXVIII. shows, it is a very small plant, with an
extremely slender creeping thallus, branched in the same
way as in the preceding species, but differing from it in
being attached to the rock by flat membranous feet, like
those of Tristicha ramosissima(Pl. XXXVII., fig. 10). Some-
times these feet project under the thallus, while sometimes
they appear to be formed by a widening of the whole
thallus ; both cases are shown. I have not seen the flower*
The fruit is slightly stalked, and is dehiscent, with a larger
upper lobe. Very commonly the latter has two or more
(54)
404 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
extra ribs, so that the fruit, which is otherwise like that of
Hydrobryum lichenoides, becomes ten- to twelve-ribbed, and
not unlike the fruit of H. Griffithii. The internal structure
of the fruit is like that of F. metzgerioides, but the large
placenta bears four (or perhaps at times three or five) seeds
instead of two.
In general, therefore, the genus may be looked upon as
closely similar to Hydrobryum. Its most interesting special
features are the endogenous branching and the reduction of
the number and increase of size of the seeds, and their
germination in situ in one species.
GENERAL SUMMARY AND DISCUSSION.
Having described the more important features of the
general morphology and life-history of the Indian forms,
we may go on to discuss them in connection with the
general conditions of life which were considered in the
introduction, and to consider their bearing on some of the
general questions of morphology, ecology, evolution, &c.*
We have already briefly outlined (p. 277) the way in
which the general life-history of these plants is connected
with the seasonal distribution of the rainfall, describing
the way in which the seeds germinate at the beginning of
the rains, giving rise to small primary axes from which the
thalli, which form the most important morphological feature
of the order, bud out. In all but Lawia the thallus is of
“ root ” nature, with endogenous secondary leafy shoots which
ultimately become floriferous, the flowers opening with
the fall of the water in the early part of the next dry
season, and shedding the seeds upon the rocks. All these
features will now be dealt with in a general way. The most
interesting general morphological points are perhaps the
* Summaries of the chief features of general importance are given after
each genus described above ; cf. pp. 303, 325, 339, 362, 369, 374, 395, 404.
OF CEYLON AND INDIA.
405
extraordinary plasticity of the skeleton-less root, and the
parallel dorsiventrality of the vegetative and floral organs,
the latter point in particular leading to important deductions
bearing on the importance of correlation in evolution.
The Primary Axis."^
We now know the construction of this organ in Lawia,
some Dicræas, Podosternon subulatus, Hydrobryum liche-
noides, H. olivaceum, Farmeria metzgerioides, and probably
Willisia selaginoides, while it is not improbable that the tall
floriferous shoot of Sphærothylax, as described by Warming
(42, IV.), is also a primary axis. It is very much to be
desired that the primary axis should be investigated in the
other Podostemaceæ, especially in some of the Tristicheæ,
as we may hope from it to get some light on the difficult
problem of the phylogeny of these plants. Looking at
such cases as the probable primary axis in Willisia and
Sphærothylax, and the well grown hypocotyledonary shoot
of Hydrobryum olivaceum, it would seem probable that the
order at first consisted of forms with well grown primary
axes, which in all likelihood were floriferous. From this
stage the evolution would appear to have pro.ceeded in the
direction of a gradual reduction of the primary axis, as the
secondaries, and afterwards the thalius also, took over its
functions of assimilation and flower-bearing. The primary
axis in Hydrobryum lichenoides, and still more in Dicræa
and Farmeria, is reduced to a very insignificant object, and
after germination is over, and the thalius established, it
appears to be of no further value in the economy, and does
not bear flowers.
The phyllotaxy of the primary shoot is usually complex,
but in Podosternon subulatus it is approximately distichous
like that of the secondary shoots. The leaves are usually
almost exactly like those of the secondaries, and are of very
simple construction, as in most water plants. The anatomy
And of, preceding paper, p. 190.
406 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
of the stem is like that of other highly adapted water plants,
but requires study in detail.
The primary axis in most forms shows little sign of
dorsiventrality, unless in the excentricity of the bundles.
In Lawia the curious lateral growth, of or from the primary
axis, resulting in the formation of the thallus, requires
further investigation. In Podostemon subulatus the primary
shoot shows a certain unequal symmetry in the way it forms
an upper side by the curving over of the leaves, while
in Farmeria metzgerioides there seems to be an actual
dorsiventrality in the very cotyledons, which are diverted a
little towards the lower side, and slightly unequal. None
of the PodoBtemaceæ as yet investigated have any developed
primary root, as indeed is only to be expected from their
mode of life.
The ThalBys and Secondar^f Sheets.
These organs are best considered together, as they are
essentially connected in the general ecology and morpho-
logical construction of these plants, and not infrequently
also, as in Lophogyne, the secondary shoots themselves are
developed in thalloid forms, making what we may perhaps
call secondary thalli. Under the general term thallus or
primary thallus we may include the creeping dorsiventral
organ which is developed from the primary axis and itself
bears endogenous secondary shoots. These, in the Indian
forms, are not themselves thalloid in structure, but are simple
leaf-bearing axes, ultimately floriferous in all but Lawia.
Phylogenetically considered, the thalli found in the different
genera of the order are of different morphological value. It
is evident that the thallus of Lawia is not homologous with
the other Indian thalli, and it is doubtful if it is exactly
equivalent to any of the other thalli found in the order. It
is evidently of shoot nature, if we judge only by the upper
side, which bears leaves, and whose growing points end in
flowers, but, on the other hand, we have to note the curious
growing point with its root-cap-like lower edge, and the
OF CEYLON AND INDIA.
407
peculiar method of formation of the whole thallus by a lateral
outgroAvth apparently of the entire primary axis, so that there
is a possibility, though probably a very unlikely one, that we
should regard the thallus as ‘‘ combined ” root and shoot, and
perhaps the endogenous shoots as representations of its
“ root ” element. The whole question, however, must be
left for detailed developmental investigation. Castelnavia
princeps has a shoot thallus, apparently derived from the
primary axis, but in this there seems to be a combination of
stem and leaf, which is not the case, so far as we can see, in
Lawia. In Lophogyne arculifera the root-thallus bears
secondary shoot-thalli, of similar morphological construction
to the thallus of Castelnavia princeps.
The other seven Indian genera all possess some kind of
‘‘ root ” thallus, and this is also the case with most of the
American genera. Taking first the Tristicheæ, Ave find in
Tristicha ramosissima and Weddellina squamulosa a thallus
to Avhich, Avithout any serious stretching of the common
meaning of the term, the name of root may Avell be applied.
It is a thin thread-like organ, which is probably endogenously
developed from the base of the primary axis just as in
Podostemon subulatus, in which a similar organ occurs. It
is endogenously branched, at least in Tristicha, and the
structure of the vascular cylinder, though slightly dor si -
ventral, as might be expected, is not markedly different from
that of an ordinary root ; the tip is covered by a root-cap of
the ordinary kind. The only marked feature in which these
roots are peculiar is the regular development upon them, in
acropetal succession, of the endogenous leafy shoots, and this,
though carried here to a high degree of perfection, is by no
means uncommon in ocher roots. These secondary shoots
reach a high degree of development and complexity, but are
not otherwise remarkable among water plants. In Tristicha
hypnoides, accordiug to Cario (8), there is no root-cap, but
otherwise it seems to resemble the other Tristicheæ.
Passing on to the more modified forms in the order, we
find the thread-like root-thallus in nearly all the other
408 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
American forms, without much difference from what has
been described in Tristicha, excepting a reduction of the size
and complexity of the secondary shoots.
In the Indian Eupodostemeæ we find the most remarkable
types of thallus in the order. In Podostemon subulatus the
thallus arises endogenously from the base of the hypocotyl,
but is itself exogenously branched, and has a very curious
collenchymatous and non-deciduous root-cap ; it is still
thread-like, but is perhaps rather more dorsiventral than
that of the American Podostemons, and its secondary shoots
are much smaller. In P. Barberi the dorsiventrality is
more marked, the thallus being broad and almost leaf-like,
though it is still attached to the rock in the middle line.
In the remaining Indian genera, there are several types
of thallus structure. Thus in Dicræa and Griffithella we
find the very remarkable alga-like forms above described,
some of which, e.g.^ the creeping closely -attached forms of
D. stylosa fucoides and G. Hookeriana Willisiana, are not
very different from the thallus of Podostemon Barberi ; they
are endogenously developed from the base of the hypocotyl,
have very reduced secondary shoots, are exogenously
branched, markedly dorsiventral in their external and
internal structure, and have the collenchymatous root-cap.
From this form it is an easy stage to the more or less freely
drifting type of thallus found in the same varieties, and to
the somewhat more modified and slightly dimorphic thalli
of D. elongata and D. dichotoma or D. stylosa laciniata.
Even the extraordinary forms found in Griffithella, such as
the stalked goblet in PI. XXYI., fig. 1, are led up to by easy
stages through intermediate forms.
In this connection attention may be specially drawn to
the remarkable polymorphism exhibited by these thalli, and
which is well shown in the figures of Plates XXI.-XXVL
The thalli of many of the other genera are very variable,
but none approach the wide range of form seen in the thalli
* Cf. 42, under Ligea, Apinagia, Mourera, Lonchostephus, Marathruui,
Podostemon, Mniopsis, &;c.
OF CEYLON AND INDIA.
409
of the Dicræas and Griffithellas. The reason perhaps lies in
the fact that the latter are usually free of the rock except at
the base, and that they possess more capacity for growth in
thickness of the cortical parts. This capacity, combined
with an entire absence of any controlling skeletal tissue,
such as is found in most flowering plants, or of fixed
position, such as hampers the thalli of the closely attached
Hydrobryums or Lawias, gives them a potentiality of varied
form unexcelled above the somewhat similarly circum-
stanced brown algæ. For details of the very simple way in
which all the very varied forms are produced, reference
must be made to the details given above under Griffithella
and Dicræa. The effects are sufficiently striking ; it is
difficult to realize that the objects figured in Plate XXVI. are
really roots, or that most of them do not even differ varietally
from one another.
Another line seems to be that represented by Hydrobryum
lichenoides and H. sessile, as well as by the Farmerias.
This form of branched simple ribbon-like thallus differs
from that of Podostemon subulatus and P. Barberi in its
greater branching, very firm attachment to the rock, and
greater flattening and dorsiventrality, with extreme reduc-
tion of the secondary shoots. In Farmeriathe branching is
again endogenous, as in the forms with which we started,
though very likely this phenomenon is one acquired subse-
quently in exogenously branched ancestors.
In H. sessile we see the thallus branching in the same
way, but wasting no space upon the rock by long gaps
between the branches, and at times there is even a tendency
for the sinuses between the lobes of the thallus to disappear
by subsequent growth. Carry this feature a stage further
and the deeply lobed liverwort-like thallus of H. Griffithii
or H. olivaceum griseum is reached, and with a further
development of the growth of the bases of the sinuses the
almost circular-outlined thallus of H. olivaceum zeylanicum
is formed. In these plants the term thallus is very
eminently suitable to the organ in question, which hag
410 WILLIS ; MORPHOLOGY OF THE PODOSTBMACB3Æ
hardly any typical ” root character left to it. It is
developed from the base of the hypocotyl, sometimes endo-
genously but often exogenously, it increases by marginal
instead of apical growth, is irregularly lobed, has a marginal
“root-cap,” and performs the function of assimilation, in
addition to bearing the floriferous shoots.
As mentioned above in dealing with Hydrobryum, if we
had only these plants before us, it would be difficult to know
what to make of these thalli, but with the series of grada-
tions before us, we must consider them as homologous Avith
the thalli of Tristicha and the American Podostemons.
These being unquestionably roots, thougli adventitious, in
the ordinary morphological sense, there is no reason to
refuse the term root to any of these thalli. On the other
hand, though this is no doubt true, if we call these thalli
roots we must use the latter expression in a very broad and
A^ague sense, and as a classificatory rather than a descriptive
term, much as we use the terms Dicotyledon, Fungus,
Flower. The term root has so much more specialized and
functional a meaning"'^ than the terms stem or leaf in
botanical work, that it is probably better not to apply it
definitely to the roots of most of these Podostemaceæ, but
to describe the organs in this family as thalli, leaving for
further specification in each case whether the thalius is of
“shoot” or “root” nature. None of the ordinary defini-
, tions of roots can be made to include the thalli of such plants
as Hydrobryum Avithout stretching them so much that it is
difficult to exclude stem forms.
This of course is assuming that these thalli are phylogene-
tically descended from undoubted roots, i,e.^ t])at some of
their ancestors Avere provided with roots developed “ adven-
titiously” at the base of the hypocotyl. There is eA^ery
reason to suppose that such Avas the case, for Ave know this
to^be an extremely common feature in many plants, and
especially in those Avhich are unable for physical reasons to
* The reasons for this have often been pointed out, e.g., by Sachs and
Goebel.
OF CEYLON AND INDIA.
411
form proper primary roots. A very large number of plants,
as is well known, show such a development of adventitious
roots, especially Monocotyledons and water plants.
The question then arises, whether an organ of “ root ”
structure (i.e., in general possessing the morphological and
anatomical features of true roots, such as the presence of cap,
absence of leaves, endogenous branching, centripetal primary
xylem, &c.), developed endogenously and laterally at the
base of the hypocotyl, is really a root. It is evident that it
cannot be regarded as strictly homologous with the true
primary root developed from the embryo, nor with roots
which arise as branches of the primary root. No one, how-
ever, denies the title root to such growths, though it is
usually qualified by the adjective “ adventitious,” and we
perhaps hardly know their phylogenetic morphological value.
It is every day becoming more clearly recognized"" that there
is much less of strict homology among the organs of plants
than has hitherto been supposed, and that just as the evolu-
tionary tree of the various groups of plants themselves is
now constantly requiring pollarding (to quote a well known
morphologist), or even perhaps coppicing, so too the organs
found among plants may have arisen by many different lines
of development, and that it ,is almost as difficult to fix the
morphological value of an organ as to fix the systematic
position of a plant, on a phylogenetic basis.
To prove the absolute root ” nature, then, of these thalli,
is by no means easy, especially as we have not yet quite*
decided what is a root. We ought, to prove the homology
of primary root and adventitious root, to trace back the
phytogeny of each, till we find the two lines unite, by one
developing from the other, or both from some parent form.
In the absence of any clear knowledge of the phytogeny of the
various groups of plants which we unite under Pteridophy ta
and Spermaphyta we cannot at present do this with any
accuracy. But so long as we regard as roots the organs
cf. Bower, Presidential Address to Botanical Section, British
Assoc., Bristol, 1898,
(55)
412 WILLIS : MORPHOLOGY OF THE PODOSTEMACBÆ
which at present go under that name in the various groups
of Pteridophyta, in the Gymnosperms, and in Monocotyledons
and Dicotyledons, and include in the category not only the
primary or embryonic roots, but also the roots which are
adventitiously developed from stems or other parts, we must
include in the general term the thalli of the Podostemaceæ.
For the present this is no doubt the best course, just as it is
still, and will probably long remain, best to continue using
such classes as Fungi, Gymnosperms, Sympetalæ, though we
now consider them as almost certainly polyphyletic ; until,
however, our knowledge of the phyla is much more com-
plete, it will be unsafe to venture far in proposing new
classifications based thereon. Let us then accept the term
root as the general classificatory expression for a group of
organs which have certain features in common, but which
have in great probability not all arisen from one phylum
of descent. In this large group we must make many small
ones, but it will be safer not to draw too many lines of con-
nection through these until we know more of their actual
phylogeny. Thus we may make classes for the roots
derived directly from the embryo in each group of Pteri-
dophyta and Gymnosperms, and probably may include in
another the corresponding roots of Monocotyledons and
Dicotyledons, the latter class probably to be united to one or
more of the former. Secondary or adventitious roots, again,
must apparently form another polyphyletic group, with many
sub-groups, among which one must be kept for the thalli of
the Podostemaceæ, another for the thorny roots of certain
palms, and so on.
Our conceptions of the morphology of the root require to
be much enlarged, just as has happened in recent years in
the case of shoots, sporophylls, &c. The root in general is so
uniform in structure, owing to general uniformity of func-
tion, and is so concealed from view and inconvenient to
work with, that its morphology and other features ha^e
tended to be neglected in favour of those of the subærial
organs.
OF CEYLON AND INDIA.
413
In those cases where the primary axis is soon lost, as in
Dicræa, Hydrobryum, or Farmeria, the mature plant forms
an interesting parallel to the case of the aquatic Utricula-
rias;^ the latter are independent ‘‘leaves” of very varied
form, the former independent “ roots.”
The formation of secondary shoots upon the roots is of
course common in other families,! but it is carried to a
degree of perfection in the Podostemaceæ that is unmatched
elsewhere. It is clear that we cannot regard the secondary
shoots of Lawia as homologous with those of the other
genera, at any rate on the existing evidence, but we may
fairly regard the secondary shoots which arise upon
unquestionable root-thalli as homologous with one another,
though they are.not of course homologous with the primary
shoots, from which they often differ very largely in structure,
as well as in position and origin. It is evident, therefore,
to make a momentary digression, that it is unsafe to draw
morphological comparisons between the Podostemaceæ and
other families, using the primary axis of the latter, and only
the secondaries of the former.
The secondary shoots of the other genera than Lawia are
borne upon the root-thalli, endogenously, and are formed near
the growing apex in alternate acropetal succession. There is
often a marked tendency for these shoots to be developed in
pairs opposite to one another, but only rarely is the oppo-
siteness found to be universal in any particular thallus.
In Tristicha the secondary shoot is large and complex,
and ultimately bears many flowers ; we do not yet know in
what respect, if any, it differs from the primary axis. The
ramuli, or branches of limited, growth, are a curious feature
of this genus and its ally W eddellina. In the South American
Oenones, Moureras, &c., the secondary shoot is large, usually
with large leaves, is more dorsiventral than in the Tristicheæ,
and bears the flowers in a large inflorescence. In the
Eupodostemeæ it is much smaller. In Willisia, to deal
* Goebel, in Ann. Buitenz., IX.
t Goebel, Organog’rapMe der Pflanzen, p. 178.
4T4 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
mainly with the Indian forms, it is still fairly large, but not
branched, and as a rule bears one terminal flower only. Its
leaves are like those of the primary shoot, but in a different
phyllotaxy, and at length they form scaly bases. The
anatomy of the stem itself is also different from that of the
primary axis, not merely in the presence of the stiffening
layer of tissue needed to hold the stem erect, but also in the
construction of the vascular tissue, which shows decided
dorsiventrality and other points of difference. In Podoste-
mon the reduction of the secondary axis has gone further,
and it is short and more or less prostrate, though branched,
and bears few flowers ; it still, however, does a major part
of the work of assimilation in the plant, as its leaves are
large. It is much larger and more complex than the primary
axis in the one species in which it has been investigated.
In the remaining four Indian genera we find an extreme
reduction of the size of the secondary shoots, accompanied
by a flattening and enlargement of the thallus, which does
much of the work of assimilation. In Hydrobryum,
Farmeria, Dicræa, and Griffithella, the secondary shoots are
at first mere tufts of leaves, endogenously formed in acropetal
succession on the root-thalli, but towards the flowering
season they elongate and bear each a few bracts, a spathe,
and a terminal flower. They are not very dorsiventral until
the flowers develop, and then there is usually an extreme
dorsiventrality, most marked in Hydrobryum and Farmeria.
In these forms the primary axis is also reduced to a great
degree.
The most interesting general evolutionary features of the
series of secondary shoots, then, are their gradual reduction
in length, branching and complexity, and the diminution of
the number of flowers to one, and finally the non-elongation
of the axis until that one flower has to be developed upon it.
The same kind of evolution seems to show in the American
forms, from the large and complex secondary shoots of the
Weddellinas and Moureras to the comparatively simple ones
of the Podostemons and Mniopsides.
OF CEYLON AND INDIA.
415
There is of course no direct evidence, other than that
afforded by a general comparison, for the view that the
small size and little complexity of the secondary shoots in
the genera named is actually due to a reduction, but it
would seem not unlikely that such has been the case.
A very noteworthy general feature in many of the thalli
is their exogenous branching, and a brief consideration of
the various cases may help to throw some light on the
meaning of the difference between the two methods of
origin of new organs, endogeny and exogeny. In the case
of an ordinary subterranean root, it is easy to see that
endogeny is advantageous to the lateral roots, and perhaps
also necessary. The extreme tip of the root is forcing its
way through the soil, and it would be difficult for either
lateral roots or hairs to form on it without injury. By the
time that the lateral roots do form, the surface cells have
become full grown and comparatively thick-walled, and it is
perhaps almost impossible for them to become again fully
meristematic in such a way as to give rise to a new organ.
It is therefore possible that there is an element of necessity
in the endogeny of the branch root, as well as of advantage.
This view is confirmed by the behaviour of adventitious
roots formed upon the stem or upon old parts of the root,
or even upon leaves ; as a rule they are also endogenous,
though it often seems to be of no advantage to them that
they should be so. Their origin is presumably determined
by heredity or by the physical conditions of the surface
tissues. Thus perhaps on the whole it comes about that
endogeny is fixed in the heredity of the higher plants as the
customary mode of origin of roots in general. The same
reasoning also applies to the case of shoots developed upon
the roots, which are also almost always endogenous.
In the Indian Podostemaceæ, we have a case of very plastic
organisms in which, while they are apparently descended
from plants with endogenous root formation and branching,
the advantages of endogeny, and apparently also the
necessity for it, have disappeared, so far at any rate as the
416 WILLIS : MORPHOLOGY OP THE PODOSTEMACEÆ
lateral rather than vertical branching is concerned. Verti-
cal organs will be more subject to scour, and perhaps to
them it may be advantageous to be endogenously formed.
We have then left, the powerful influence of heredity, and
the time of appearance of the new organs, whether at the
earliest stages near the apex, or later. In the case of the
secondary shoots upon the roots, even when they are formed
near to the tip, endogeny is the rule without exception.
That there is perhaps some very strong reason for this may
be inferred also from the fact that even in Lawia, with its
shoot thallus, the secondary shoots are formed endogenously,
on the mature parts of the thallus, though perhaps their
late origin, when the surface cells are hard and full of
silica, may be enough to account for this. Beijerinck^ has
shown that in Aristolochia Clematitis the secondary shoots
show endogenous development when formed on old parts of
the root, exogenous when on young ones.
In the branching of the more modified root-thalli, on
the other hand, endogeny is distinctly rare ; it occurs in
Farmeria, but here the branches are formed very late in the
development. It also occurs in Dicræa when branches are
formed later in life on old parts of the root, but never at the
apex, unless in case of rejuvenescence from an injured tip.
In Tristicha the branching is endogenous, but it occurs at
some distance from the tip. In. most of these forms, how-
ever, the branching occurs quite close to the tip of the
thallus, where the tissue is still fully meristematic, and is
exogenous.
It is evident, then, that like other morphological features,
endogeny and exogeny, fixed though they are so long as
there is no serious change in function or circumstances,
may be modified in accordance with changes in these
conditions. Perhaps we may regard as one of the determin-
ing causes in the first differentiation of these two modes of
* Beob. u. Betr. ii. Wnrzelknospen u. Nebenwurzeln. Verb. k. Akad.,
Amsterdam, 1886, d. 25,-
OF CEYLON AND INDIA.
417
origin the direct advantage of endogeny in many cases, and
as another the physical necessity of it when the organ was
developed at so late a period that the surface cells were no
longer capable of full meristematic activity.
Having now traced the morphology of the thallus through
the very remarkable series of forms that it exhibits, we
must go on to deal with it from an ecological point of
view, and endeavour to trace the correlation between the
morphology and the general conditions of life, or, in other
words, to determine the amount and degree of adaptation
displayed.
One feature of great importance, common to all the
thalli, is as we have seen the development of regular
secondary shoots. In the simpler thalli, such as those of
Tristicha and the American Podostemons, or even P.
subulatus, this seems to be the only special function of the
thallus as compared with an ordinary adventitious root,
because the long creeping thread-like organ, with its
comparatively slight hold of the rock, is not suited for
purposes of anchorage of the long and complex secondary
shoots. Anchorage of the latter, as soon as they have
reached any considerable size, is provided for by the
development of haptera.
The function of the thallus seems rather to be to spread
rapidly over the rocks, developing new shoots at short
»
distances, and holding them until they can form holdfasts
of their own. Referring back to the general conditions of
life, described in the introduction, we can easily see the
advantages of this. If the water become shallow, as may
easily occur at any time by a deficiency in the rainfall, the
shoots are liable to be killed, or at any rate much injured,
by exposure to the air, whereas the creeping thallus, which
is also often more or less amphibious (or rather, capable of
standing some exposure, for it never of itself emerges into
the air), stands the best chance of surviving until a rise of
water again occurs, and then forming new shoots. At the
same time, the more shoots there are, the better for the
418 WILLIS : MORPHOLOGY OF THE PODOSTBMACBÆ
chance of the plant as a whole, for some of them will have a
chance of living, or even may never be exposed, and thus the
life of the whole may be comparatively less injured than if
there were only one or two shoots.
Another advantage of the thallus, and especially when
combined with a dwarfing of the secondary shoots, is that it
enables the plant to live in swiftly running water without
danger of being carried away. At the same time, the
importance of this advantage may be easily exaggerated.
Hydrobryum olivaceum, the species which inhabits the
swiftest water of all, has a tall primary axis with long leaves
at the period when the water is deepest and most violent,
and this axis can hold fast to the rock if the rest of the
thallus be taken away. Many of the South American forms
have very large leaves and shoots, and yet live in very swift
water. We must therefore beware of regarding the thallus
too much as an adaptation to life in swiftly moving water.
A third factor to be considered is the scour of the sus-
pended matter in the water, which will be greater where
the current is swifter, and which will probably be more
injurious to the larger forms.
It would seem, then, on the whole, most reasonable
to regard the thallus with its secondary shoots, the chief
morphological feature of most of the Podostemaceæ, as pro-
bably adapted less to the rushing water as such than to the
dangers inseparable from life in such water, chiefly the
ever present risk of exposure by the shallowing of the water,
but also perhaps the increased scour. The more highly
modified the thallus, the swifter on the whole is the water
in which it lives, as we shall see in detail in dealing with
the geographical distribution of the forms. We may also,
perhaps, regard as adapted to the same dangers the slight
amphibiousness of most of the thalli (perhaps partly due
to the silica that they so often contain), and their great
capacity for rejuvenescence by forming new growing points
behind any injured part, or even on older uninjured parts.
OF CEYLON AND INDIA.
419
The thallus, then, appears to be well adapted to such a
mode of life, but we are now met as usual with the question
whether the adaptation is a direct one, or only a perfecting
of characteristics already existing in the ancestral forms, and
which perhaps enabled them in the first place to adopt this
peculiar mode of life. Here, as usual, we are checked in
our inquiry by our ignorance of the phylogeny, and still
more in this particular case by the absence of other families
of flowering plants living in similar conditions, with which
comparisons can be made. Neither the development of
adventitious shoots on the root nor the regeneration of the
growing points is at all uncommon in other families, and
hence it seems very probable that both of these characters
were well marked in the ancestral forms of the Podostema-
ceæ. But they are evidently of great value to these plants,
and have become developed to a great degree of perfection,
and may thus perhaps best be regarded as partial rather
than absolute adaptations to the mode of life.
While the thallus, regarded broadly, is thus apparently
not a direct adaptation, it is by no means improbable that
there may be in it many partial direct adaptations, and we
must now trace it a step further into the more peculiar forms,
such as the Dicræas and Hydrobryums. In these we begin
to find forms unique among the higher plants, and may
therefore look for more evidence of direct adaptation.
We have seen that the peculiar expansion of the thallus
so marked in these highly modified forms is accompanied
by a dwarfing of the secondary shoots, and perhaps on the
whole by a slight increase in their numbers. It is easy to see
that the dwarfing is in some ways advantageous to the plant,
as it decreases the risk of exposure to the air when the water
becomes shallow, but, on the other hand, it also decreases the
assimilatory area and capacity of the plant, and the number
of flowers borne on each shoot. Increase in the number of
the shoots compensates more or less for these losses in some
cases, and also such features as the development of larger
leaves in Podostemon subu latus or Hvdrobryum olivaceum,
(56)
420 WILLIS : MORPHOLOGY OF THE PODOSTEMACBÆ
but at length in the cases of extreme dwarfing seen in
Dicræa, Griffithella, Hydrobryum, and Farmeria, we find
the expanded green thallus the chief organ of assimilation,
taking on more or less of the anatomical and morphological
construction of a leaf, or rather of the flat thallus of many
Bryophyta or Algæ. These two features, the dwarfing of
the shoots and the expansion of the thallus, are evidently
more or less correlated, but we are met by the usual diffi-
culty, whether one or the other, or neither, may be regarded
as cause or effect.
The peculiar thalli just mentioned may be roughly divided
into two groups, the algal forms of Dicræa and Griffithella
(and some Podostemons to a less extent), and the more lichen-
or liverwort-like forms of Hydrobryum and Farmeria. The
extraordinary resemblance of many of these plants to algæ,
especially the seaweeds of the rapidly moving or rising and
falling water of the seabeach, has often been noticed, and
some special cases of resemblance in detail have been pointed
out above. Comparison of the general conditions of life
given in the introduction with those which obtain for sea-
weeds in such positions as are occupied by many Fuci,
Himanthalia, Bostrychia, and many more, will shoTv that
there is a considerable agreement in the two cases. That
this is so is also suggested by the behaviour of the interesting
algæ described by Goebel,'" found at the tidal mouths of large
rivers (but in fresh, not brackish water, owing to the great
size of these streams), where they are exposed to the air
during part of the day, Bostrychia Moritziana is one of these
forms, which in its morphological construction is closely
analogous to Podostemon subulatus, and which also occurs
in the rapids of the mountain streams of Guiana, occupying
there the same localities as the Podostemacea, Oenone
Imthurni, a plant with filamentous creeping roots, upon which
are borne prostrate secondary thalli of shoot nature, but not
* lieber einige Süsswasserflorideen ans Britiscli-Guyana ; Flora, 83,
1897, p. 436 ; Eine Süsswasserfloridee aus Ostafrika ; Flora, 85, 1898, p. 65 ;
and Organographie der Pflanzen, pp. 30-34, &c.
OF CEYLON AND INDIA.
421
unlike in form to the leaf-like root-thalli of Dicræa Wallichii
or some forms of D. stylosa. Another of these algæ, Deles-
seria Leprieurii, is remarkably like Hydrobryum lichenoides
or Farmeria metzgerioides, leaving out of consideration the
comparatively insignificant secondary shoots of these forms.
The leaf-like thallus of some of the Dicræas, again, is
parallelled not only by the shoot thalli of Oenone and other
South American forms, but also by the large leaves of many
American forms, and all of these organs are in many respects
closely similar to the thalli of many of the seaweeds. Still
more curious forms of the Podostemaceous thallus occur as
we have seen in Griffithella, but these are also parallelled in
the Himanthalias and other algæ. The whole subject of
these resemblances between plants far distant in relationship
from one another is of course one of great difficulty, and it
is impossible at present to do more than point out these very
suggestive analogies of form which accompany analogy of
the conditions of life, and which seem to indicate that an
experimental and comparative morphological study of the
forms of the Algæ and Podostemaceæ should be attended
with interesting results. The parallellism also indicates that
it is by no means unlikely that the extraordinary poly-
morphism of the thallus, which occurs in certain genera and
species of the Podostemaceæ, e.g.^ in Dicræa and Griffithella,
may have its parallel among the algæ, and consequently that
the latter group requires investigation from this point of
view, to strengthen its taxonomic foundations. In both cases
we are dealing with organs which are not hindered by a
skeletal tissue from growing into almost any shape, and in
which therefore a variety of form may be easily liable
to occur with any slight change in the conditions of
growth.
Not merely do many of the Podostemaceæ present close
similarities to the seaweeds, but the least modified group, the
Tristicheæ, presents remarkable similarities in morphological
features, and in the arrangement and anatomy of the leaves,
to many mosses or liveifworts, especially to those of wet
422 WILLIS : MORPHOLOGY OP THE PODOSTEMACEÆ
situations. Here again, comparative and experimental
morphological study is required in both groups.
To return from this digression, it would thus seem probable
that the peculiar more or less expanded and assimilatory
forms of the thalli (whether “shoot” or “root”) in these
highly modified plants, unique as they are among the higher
plants, and parallelled in the lower plants Avith more or less
similar conditions of life, taken together with the reduction
of the secondary shoots, are largely adaptations to the mode
of life, more especially to the rising and falling water with
the attendant risks of exposure. The thalli and secondary
shoots of course shoAAq as Warming and Goebel have already
pointed out, the common adaptations to life under water,
such as flexibility of the drifting parts, absence of water-
carrying or storing tissue except at the period of exposure
for flowering, and so on. The most marked peculiarity, the
absence of the large intercellular spaces, has been dealt with
under Tristicha (p. 305), where it was mentioned that Goebel’s
explanation is probably the correct one, viz., that as the water
is fully aerated, there is no need for the air spaces which in
many water plants are needed for the oxygenation of the
lower parts. That the Podostemaceæ are very closely
dependent on the motion or aeration of the water is shown
by the fact that if placed in still Avater in a vessel, or if cut
off in a pothole by the fall of the level of the river, they
quickly die.
Finally, we must briefly refer to the interesting stages
shown in the life-history of the thalli, especially the changes
that take place at the flowering season. In Ceylon they reach
their full vegetative growth by about the end of October, and
the secondary shoots begin to form their flowers in the last
months of the year. By the time Avhen the level of the water
falls at the end of December the flowers are quite ready to
open. Further north of course the floral development is
earlier, corresponding to the shorter rainy season. In
Tristicha ramosissima the flowers are developed upon all or
OF CBYLOîf AND INDIA.
423
most of the secondary shoots, upon short branches, and
emerge from the water as soon as its level is sufficiently low ;
by the time that the seeds ripen the thalli are commonly
quite exposed upon the rock. In Podostemon subulatus the
process is similar, but the flowers do not open till the water
falls low enough to expose the tips of the spathes to the air.
In the other genera the phenomena are more specialized, the
secondary shoots being reduced and single-flowered. In
general, all the secondary shoots do not become floriferous.
In Hydrobryum, Griffithella, andFarmeria, it is usually the
apical shoots which remain vegetative, and the same is the
case in most of the Dicræas, but in D.' Wallichii and in some
forms of D. stylosa, and also at times in Griffithella, scatter-
ed shoots here and there over the whole thallus remain
vegetative, while the rest flower. Details must be sought
under the descriptions of the genera ; the interesting general
point, and one which {ef. preceding paper, p. 191) has
important taxonomic bearings, and has in the past led to
■much error and confusion, is the curious way in which the
non-floriferous parts of the thalli fall away, leaving the now
woody flowering portions.
During the flower development in the forms with reduced
secondary shoots, the axis elongates, and an interesting
change of form and function takes place in the leaves which
are formed at this time (for details see Dicræa, p. 350), their
sheathing bases enlarging to form bracts, while the filamentous
assimilating tips finally fall away ; this process also, as
explained in the preceding paper (p. 193) has led to many
erroneous conceptions in taxonomic work based on her-
barium material. Once fully exposed to the air in the
flowering period, these tips soon fall off, together with any
not woody parts of the thalli, so that the whole appearance
of the plant is often changed. This is well illustrated in
the plates accompanying this paper {e.g., cf. XIV. with XV.,
XIX., 1 and 2, XXXIII. with XXXV., XXXVIII., 6, with
V., &c.).
424 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
The Leaves»
'These require but little discussion here. A distinction
must of course be indicated between those of the primary
and those of the secondary shoots, though in fact they appear
to be quite similar in structure. In the Tristicheæ the leaves
are singularly like those of many mosses, small, entire, and
extremely delicate. In Mourera and other South American
forms they are often of large or even immense size, and
exhibit forms like those of many marine algæ. In the Indian
Podostemeæ, however, they are always simple and usually
subulate or linear. The two species of Podostemon, and
Willisia selaginoides, have long leaves of considerable size,
and those of Hydrobryum olivaceum may reach 10 cm. in
length, but in the other forms the leaves are very small,
rarely over 1 cm. long.
The dimorphism of the apical leaves in Lawia is very
interesting, and is parallelled by what is found in Selaginella,
and analogous to the construction of some mosses or liver-,
worts. A time-dimorphism occurs in nearly all the other
genera found in India. The vegetative leaves, as already
sufficiently described, have slightly sheathing bases, but
while the flower is developing the bases of the leaves become
much enlarged and form sheathing scaly bracts, while the
tips fall away. These scales show a lateral dorsiventrality
of structure, in that the sheath is thicker on the upper side,
and the more so the more marked the dorsiventrality of the
secondary shoot.
A noteworthy feature of the leaves of the Indian Podos-
temeæ is the development of hairs upon the upper side,
presumably for enlarging the absorbing area, as with the
“ Kiemenbüschel ” described by Goebel and Warming in
many South American forms.
The Hapfera.
The morphology and the phylogeny of the peculiar
anchorage organs which are so common in the order, and
OF CEYLON AND INDIA.
425
which Warming has named haptera, are questions of some
difficulty. They are in general formed exogenously from
the surface tissues of thallus or shoot, while these are young,
and grow by a growing point rather like that of the roots.
Not infrequently the apex shows a superficial layer of cells
like the collenchymatous cap which occurs in most of the
root-thalli, but very slightly marked, somewhat like that of
the thallus of Podostemon subulatus. Reaching the sub-
stratum the haptera flatten out upon it. The flattening, and
still more commonly a branching of the tip, commonly begin
before the hapteron has reached the rock. As described in
detail under Tristicha (p. 295), the hapteron behaves like a
large root-hair or rather rhizoid.
Haptera, though exceedingly common, seem only to be
developed when there is a definite need for their services as
holdfasts, and it is consequently difficult to determine what
is their phylogenetic morphological value, or whether indeed
their occurrence is not another expression of the great
plasticity of the tissues in these plants. Similar organs are
common enough in Algæ, Lichens,* &c. Warming and
Goebel regard them as organs sui generis^ and probably
this is the best course, though there seems no absolute
improbability in their being modified adventitious roots,
which have already gone through the change from endoge-
nous origin to exogenous, such as seems in progress in the
thalli of Hydrobryum olivaceum, &c.
The Spathe and Pedicel, &c.
The formation of the floral shoot from the vegetative, with
the interesting change of structure and function in the
leaves, has been sufficiently described. The interest of the
floral bracts, spathe, and pedicel is chiefly taxonomic, and
has been discussed from this point of view on p. 193 of
the jjföceding paper. It is of particular interest in the
present connection to point ouc the gradually increasing
*6/. Sernander in Bot. Notiser, 1901 ; abstr. in Bot. Centr,, 88, 293.
426 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
dorsiventrality of structure of these organs which accom-
panies the increasing prostrateness of the shoot. In the
comparatively erect shoots the bracts are only slightly thicker
on the upper than on the lower side, and the spathe stands
erect and splits fairly uniformly into teeth at the tip, while
in the Hydrobryums, &c., the bracts are very prostrate and
much thicker on the upper side, and the spathe lies prostrate,
splitting on the upper side to let the flower escape. In
Lawia the spathe is physiologically represented by the
cupule, which shows a marked dorsiventrality similar to that
of the vegetative growing apices, and varying in the different
forms.
Another interesting feature about these organs is the
deciduous cortex of the pedicel, a phenomenon common
to all the Indian and probably most or all of the other
species. The length of the pedicel, as pointed out in the
preceding paper, is a very variable feature, and the causes
determining it require investigation.
The Flower and Fruit.
The main features of the morphology of these organs have
been considered from a taxonomic point of view in the
previous paper (p. 194), and need not be repeated here.
Detailed accounts of many flowers and fruits may be found
in the works of Tulasne and Warming, and also in the
present paper.
The most interesting general features of the floral morpho-
logy are the substitution of the spathe for the perianth in
all the more modified tribes, and the progressive dorsiven-
trality of the flower, which will be considered below. This
dorsiventrality at last shows in the fruit also, the lower
loculus being reduced or even abortive, and the ripe fruit
splitting obliquely into tAvo unequal valves. The develop-
ment or absence of ribs in the wall of the ripe fruit is a
character of anatomical and taxonomic importance. It has
been sufficiently described and figured above under Dicræa,
OF CEYLON AND INDIA.
427
Podostemon, and Griffithella, which are good examples of
the isolobous and anisolobons ribbed, and the smooth fruits
respectively.
The natural history or ecology of the flower, like so many
other features, also affords an interesting comparative study.
The American Tristicheæ have a conspicuous entomophilous
flower, projecting from the water on a long stalk, radial in
structure, provided with a perianth, and freely visited by
insects, to judge from the authors quoted by Goebel (13, p.
330). So far as the herbarium specimens show, most of the
flowers seem to set seed, and as the seeds are small and very
numerous, the plant has probably every advantage needed
in this respect. In the next groups we find the perianth
replaced by the spathe, an organ apparently composed of one
or more of the uppermost leaves of the secondary shoot. As
to how the change occurred, and as to whether it was of any
advantage to the plants, we are as yet ignorant ; further
study is needed in South America. The perianth in these
forms, e.g.^ in Oenone, is still represented by a ring of small
scales alternating with the andrœceum ; the flowers are still
insect- visited and emerge as before above the water, but as we
progress along the series towards the Eupodostemeæ, we find
a reduction of the size and conspicuousness of the inflo-
rescence and of the individual flowers becoming more and
more marked, and the flowers becoming dependent on the
wind or on self-fertilization. At the same time the dorsi-
ventrality of the flowers is increased by the loss of the upper
stamens and perianth.
In the Eupodostemeæ, finally, we find a most remarkable
state of things. The flowers are small and very dorsi-
ventral, the perianth being absent or represented only by
the two thread-like organs at the sides of the andrœceum,
which itself is reduced to a single or forked stamen on the
lower side of the flower. Even the fruit is dorsiventral in
structure in some forms.
Dorsiventrality is usually supposed to be an attribute of
entomophilous flowers, and tobe a feature of direct advantage
(57j
428 WILLIS : MORPHOLOGY OF THE PODOSTEMACBÆ
to them, and indeed this is probably very often the case.
Here, however, we meet with the most extreme known
zygomorphism in a group the flowers of which are wind-
er self-fertilized, stand erect, and are very inconspicuous
and not visited by insects. The dorsiventrality of the
flowers seems, as will be discussed below in detail, to have
been acquired at the same time that the flowers were pro-
gressing in the direction of increased anemophily and
autogamy, and appears to be of absolutely no use to them,
except in so far as it represents a reduction of the amount
of material expended upon the construction of the floral
organs. The more dorsi ventral the flowers the more de-
graded their type, on the whole ; they have, as compared
with the higher American types, lost much of the certainty
of the cross-fertilization, and gained little or nothing in
certainty and abundance of seed, but only in reduction of
material. However dorsiventral the flower becomes it still
stands erect as long as it possesses a stalk, and when at last
we come to the forms without the stalk we find the flower
curving its ovary and stamens so as to get them as erect
as possible. It seems as if the flower were, so to speak,
struggling against the dorsiventrality to the last, and doing
its best to try for a cross, even when the chance of such a
thing has sunk to a very small figure.
It is difficult to resist the conclusion in looking at these
flowers that the dorsiventrality which is the most marked
feature of the order as a whole is not only no advantage to
them, but is an actual disadvantage, against which they
struggle, making various compromises, but always tending
more and more in the direction of regular autogamy, which
at last is adopted outright in Podostemon Barberi and
possibly in other forms.
In the fruits of these plants we get perhaps even a clearer
illustration of the general principle now under considera-
tion. The ifruit in this order is evidently a “land” fruit,
and seems quite unsuited in itself or in the seeds for an
aquatic existence ; it shows no such adaptations as we may
OF CEYLON AND INDIA.
429
find in the N^^mphæaceæ or other orders of water plants.
It opens only in comparatively dry air, and the seeds ai*e
adapted to land existence for some time before germina-
tion. They ultimately germinate under water, but they are
not suited to being shed in the first place into water. If
dropped into water from the capsule they are carried away
down stream, and have almost no chance or prospect of
becoming attached in a place suitable for growth. One is
surprised on the whole, in dealing with this order, to find
so little provision for germination of the seeds in suitable
places, where they can live and fiourish. The only advan-
tageous quality they seem to have in this respect is their
small size, which helps them to cling in crevices of the rock
or of the old thallus, but we must probably regard this as
derived from their land ancestors, not as an adaptation.
The mucilaginous outer coat, again, is rather a character of
a land plant, and probably survives in the Podostemaceæ
without being of any particular value to them in connec-
tion with their mode of life, though indirectly it must be of
service. Wading birds may often be seen walking with
wet feet over the fruiting thalli, and probably sometimes
carry seeds to other localities. When a seed falls into a crack
in an old thallus the swelling and subsequent shrinkage
perhaps help to drive it well in. The fruit is very uniform
throughout the order, the bicarpelled many-seeded symmetri-
cal form persisting from Weddellina down to many of the
Eupodostemeæ, e.g.^ Dicræa. Here at last, however, the
progressive dorsiventrality of the plants seems to show also
in the fruit, and in Podostemon and Griffithella, Willisia,
Hydrobryum, and Farmeria, the fruit is structurally dor-
siventral, with unequal lobes. As Dicræa and Podostemon
and Hydrobryum all live in similar circumstances, and all
simply shed their seeds upon the rock as soon as the erect
fruit opens, we cannot regard the asymmetrical develop-
ment of the fruit as of any direct advantage ; it would
appear rather to be an expression of the continually increas-
ing dorsiventrality of the floral parts in correlation with
430 WILLIS : MORPHOLOGY OF THE PODOSTEMAOEÆ
that of the vegetative organs. Perhaps in these cases, the
dorsiventrality of the vegetative organs being greater, the
gynæceiim comes under its influence at an earlier stage in
the development at the growing point of the secondary
shoot. By a still further increase of the dorsiventrality
we come to Farmeria indica, in which the interior of the
fruit for the flrst time shows extreme vertical asymmetry,
the lower loculus being aborted, while the placenta is
developed chiefly at the distal end. In this form, too, for
the first time, we find a reduction of the number of seeds to
four without any intermediate stages showing in any other
species, and a corresponding increase in their size. As
however they are still shed upon the rocks, it is not easy to
see if this is any advantage to them, or whether it is only
another expression of the increasing dorsiventrality ; but in
the next species, Farmeria metzgerioides, we see the (so to
speak) obvious next step taken, and the fruit, now dorsiven-
tral to the extreme degree, remaining closed and sessile
among the bracts, which thus hold the two seeds in position
for germination in situ upon the rock when the water rises.
Here again, then, the dorsiventrality seems impressed
upon the gynæceum and fruit quite apart from any
consideration of direct advantage to the fruit or to the
performance of its functions. In Podostemon the dorsi-
ventrality is still slight, it increases through the series of
Hydrobryums, and at last affects the interior of the fruit in
one of the Farmerias, but without being of any apparent
advantage. Then suddenly, we come upon Farmeria
metzgerioides, in which, so to express it, the plant has been
able at last to make a good use of the phenomenon.
The Seedlings.
It is not necessary to add much to what has already been
said about these. We now have a general knowledge of the
seedlings of Lawia, Dicræa, Podostemon subulatus, two
species of Hydrobryum, and one of Farmeria, besides a
OP CEYLON AND INDIA.
431
couple of South American forms. It is very greatly to be
desired that the other Podostemaceæ should be studied in
this respect.
In plants which when mature show very unusual features
in their morphological construction, such as the Cacti,
Utricularias, &c., we are accustomed to derive valuable
evidence as to the morphology of the plant from a study of
its seedling, and the descriptions above given of the
seedlings of many Podostemaceæ show that the seedling is
a valuable subject for study here also, throwing considerable
light on the morphology and the phylogeny of the thalli.
Most of the interesting features of the seedlings have already
been dealt with above under other heads, e.g.^ the progressive
reduction of the primary axis, the development of the
thallus, the lack of exact homology between the primary
axis and the secondaries, and so on. Dorsiventrality does
not appear in the seedlings except in one or two cases, e.^.,
in Podostemon to a very slight (and doubtfully hereditary)
degree, and in Farmeria, in which as we have already seen
the flower and fruit are extremely dorsiventral. The very
peculiar case of the seedling of Lawia, in which dorsiven-
trality is acquired soon after germination, deserves special
mention. In the other seedlings complete dorsiventrality
is not as a rule acquired, the primary axis remaining radial
in symmetry, and the dorsiventrality showing in the thallus.
Like the mature plants, the seedlings of the Ceylon forms
have simple leaves, while the American forms mostly
appear to have compound leaves appearing at once after the
cotyledons. It would be interesting to know the reason of
this difference.
The seedling plants are well adapted to their peculiar
conditions of germination and life. Except in Farmeria the
seed is very small, and the seedling also, so that it is less
liable to be carried away by the water, Haptera and
rhizoids are formed at once on germination, so that the
plant is firmly anchored, and as soon as possible the develop-
ment of the thallus begins. In Farmeria metzgerioides, we
432 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
have a peculiarly interesting case, the two seeds remaining
inside the fruit, which is held firmly against the rock, so
that germination takes place under the most favourable
conditions for the successful growth of the young plant.
The dorsiventrality, dwarfing, &c., of the mature plants
seem to show correspondingly in the seedlings, those
showing them most early which have them most marked
when mature, but we require a detailed comparative study
of the family in this respect, such as Ganong has given to
the Cactaceæ.*
Anatomy.
A number of anatomical features have been mentioned or
described in the detailed part of the paper. Much also has
been described by Warming, but there is yet need for much
comparative investigation of these very curious plants,
owing to the unusual features presented by the anatomy, in
consequence of the far-reaching dorsiventrality of the vege-
tative organs. Attention may be called to the structure of
the thalli of Lawia, Podostemon, Hydrobryum, Dicræa, &c.,
described above, to the presence of silica bodies, and to the
peculiar growth in thickness of the vascular tissues in the
thallus of Dicræa and of the cortex of the thallus in the
same genus, as well as to the peculiar root-caps, especially
the remarkable rim-like cap of Hydrobryum olivaceum.
In Dicræa and to a less extent in other genera, considerable
anatomical changes take place at the time when the thallus
is commencing to develop the flowers, owing to the
demands made by the floral shoots for supplies of water and
materials, especially when they are exposed to the air and
begin to ripen their fruits. It is hoped to follow the present
paper by others dealing with anatomy and other interesting
features of many of the genera.
Owing to the peculiar habitat of these plants, and the
rigidity given to the creeping thalli by the presence of so
much silica, it would seem not unlikely that such genera as
* Ann. of Bot., 1898, p. 423.
OP CEYLON AND INDIA.
433
Hydrobryum (cf. Plates XXXI. , XXXITI.) would readily
form fossils, which, if they had lost, as they probably
would lose, their secondary shoots, might even be mistaken
for liverworts. No fossil Podostemaceæ have as yet been
described, but it is not unlikely that some may ultimately
be found, perhaps in the mountains of North America or
North-East Asia.
Physiology.
Cinder this head a few miscellaneous notes may be
summed up, but the order still requires, and should repay,
detailed physiological investigation on the spot.
Nutrition probably takes place through the surface cells
in the same way as in other water plants, the rhizoids
appearing to be chiefly or entirely anchorage organs. An
interesting feature in the nutrition is the great quantity of
silica often present in the surface cells of the thalli, less
often in the leaves or shoots. In such thalli as that of Lawia
it is present in such quantity as to blunt a razor in cutting a
single section. Whether this silica is taken up from the
rock or from the water must at present remain an open
question, also the further question whether it has any
physiological function. It makes the thalli more rigid, and
probably helps in preserving them from too rapid drying up
when exposed to air. Those thalli which contain a large
quantity of it are able to withstand considerable exposure
and yet revive when again submerged.
These plants expose large assimilating and absorbing
surfaces, often further increased by the development of
hairs on the surface of the leaves. The leaves show the
general structure of shade leaves, and have usually chloro-
phyll in the epidermis. Most of the forms have a great
development of anthocyan in the surface cells, and
consequently a red colour when alive.
During the greater part of their life the Ceylon and Indian
forms store up vast quantities of starch and other reserves in
the thalli and bases of the secondary shoots, so that when the
434 WILLIS : MORPHOLOGY OF THE PODOSTKMACEÆ
time comes, the development of the flower, its expansion, and
the ripening of the fruit take place with great rapidity.
As a rule some starch is left in the thallus after the seeds
are ripe, and this helps in regeneration of the growing points
if it be again submerged.
As regards sensitiveness to stimuli, there is as yet little
to be said. The thalli do not seem at all rheotropic, but
the creeping forms follow out every irregularity of the sub-
stratum, whether owing to effects of contact or of light or
gravity. At the flowering period the stems change their
sensitiveness and stand erect, though while developing they
were commonly horizontal.
On the Dorsiventralify of the Podostemaceæ,
and its bearing on current views of the
Mechanism of Evolution.
No other family above the liverworts shows so marked
and far-reaching a dorsiventrality in organization, and the
Podostemaceæ are often quoted in discussions of this pheno-
menon. Their dorsiventrality has been described in
general terms by Warming (43), Goebel (13, 14), and others,
but several interesting points have as yet escaped notice,
especially the remarkable series of stages in a progressive
dorsiventrality exhibited by the members of this family,
and the curious feature presented by so many of them, of
erect anemophilous autogamous flowers, which at the same
time possess the most extreme structural dorsiventrality
which occurs among the flowering plants.
We may best illustrate this by a series of actual instances,
working downwards from the less modified types of the
order. In Tristicha ramosissima and in Weddellina squa-
mulosa there is but little structural dorsiventrality of the
plant taken as a whole ; it is only expressed in the creeping
position of the thallus. The secondary shoots stand up in
the water, and are radially constructed, while the radially
symmetrical flowers emerge through the water on erect
OF CEYLON AND INDIA.
43f)
In \¥eddellina they are entomophilous, in Tristiclia
ramosissima aneinophilous. Perhaps entomophily may be
regarded as the more primitive condition in the order, but
there is little evidence to prove this, and the point does
not matter to the present argument. In the other species of
Tristicha there is rather more dorsiventrality in the vegeta-
tive organs ; the thallus is similar to that in the preceding
forms, but the shoots arising from it show a distinct bilater-
ality, the leaves being in lateral and upper ranks, and the
branches in one plane, much as in a Selaginella or in
the dorsiventral branches of many flowering plants. The
flowers still appear to stand erect and to emerge through
the water, but they also show signs of zygomorphism, in
that there is a suppression of the upper stamen and of one
lateral stamen as compared with the flower of T. ramo-
sissima.
From these genera we branch off at once to Lawia, in
which the vegetative organs already show the most extreme
possible structural dorsiventrality, with a growing point
symmetrical only about a vertical plane, and with upward
displacement of the leaves. Nevertheless, perhaps because
this dorsiventrality is of recent acquisition and heredity is
strong, the flower is radial in symmetry, and stands erect
as soon as it emerges from the cupule. This latter organ
shows a strongly marked difference between its upper and
lower sides, but though the flower is developed inside the
cupule in a horizontal position, it does not itself show any
sign of zygomorphism, unless it be in minor anatomical
details.
The next stage is seen in the Oenones and Marathrums.
Here the root is very like that in the preceding forms, but
perhaps with rather more structural dorsiventrality, while
the secondary shoots are strongly dorsiventral and prostrate.
The flowers, however, are still erect and radial, and are
apparently developed in a more or less erect position in the
inflorescences. They are apparently insect-fertilized, but
have the perianth reduced and replaced by the spathe as
436 WILLIS : MORPHOLOGY OF THF FODOSTBMACEÆ
already described. Passing on a stage further, Ave come to
the Lophogynes and Apinagias, which have strongly marked
dorsiventrality in the vegetative organs, especially in the
secondary shoots. Here, as in Tristicha hypnoides, we find
the flowers again showing zygomorphism, in that the
stamens on the upper side are suppressed, as compared with
the radial flowers, and the upper stigma is often smaller
than the lower.
From these we pass to the Eupodostemeæ, all of which
show a very high degree of dorsiventrality in the vegetatiA^e
organs and also in the floral, but the rule indicated above
seems to hold throughout, that the dorsiventrality is most
marked in the vegetative organs, and shows less or later in
the floral shoot, spathe, stamens, stigmas, ovary, and fruit,
in the order named. Thus in Dicræa there is a highly
dorsiventral thallus, the growing point, as well as the other
morphological and anatomical features, showing the asym-
metry about the horizontal plane, but the secondary shoots,
though reduced, still stand comparatively erect until late in
their life, so that the earliest stages in the floral develop-
ment, or the impulses to that development, perhaps occur
at a time Avhen the shoot is still not prostrate. The flower
is extremely dorsiventral, and the upper stigma is smaller
than the lower, but the fruit is isolobous. The flower,
though horizontal or nearly so in the ripe bud, bends
upwards and stands erect as soon as the spathe opens.
The bracts and spathe both exhibit differences of structure
betAveen their upper and lower sides, but much less marked
than in some of the succeeding genera ; the difference is
most marked in the most prostrate shoots, 0.^., in D. stylosa.
In Podostemon and Mniopsis the general characters are
much the same, but the secondary shoots are longer and
with marked structural dorsiventrality from the very first.
The flowers emerge in a comparatively erect position, but
show more zygomorphism than in Dicræa, in that the
ovary is often rather larger in the upper than in the lower
loculus, Avhile the fruit, though erect, is ahvays anisolobous.
OF CEYLON AND INDIA.
437
the larger upper valve persisting, while the smaller lower
one falls off.
In Griffithella the vegetative organs are like those of
Dicræa, but the secondary shoots appear to be more prostrate
at an early period ; the question, however, requires further
investigation. The flower and fruit show similar zygomor-
phism to that of Podostemon. The same is the case in
Willisia, though the secondary shoots at the flowering time
are stiffly erect ; their condition in the earlier period of the
life-history has yet to be discovered. The thallus of this
species, at any rate, is highly dorsiventral. In Castelnavia
there is a highly dorsiventral thallus, accompanied by
extreme zygomorphism of the flower and fruit ; in some
species the flowers are developed in a horizontal position
in cavities of the thallus.
In Hydrobryum the thallus is also highly dorsiventral ;
the secondary shoots are like those of Dicræa, and late in life
as the flowers develop, become exceedingly dorsiventral in
structure, showing it in a high degree in the bracts, and in
the spathe, which splits along the upper side. The flower,
though developed in a more or less prostrate position,
stands erect as soon as it emerges from the spathe, at least
in those forms in which it possesses a stalk, while in the
stalkless forms the ovary and stamens curve upwards as
much as possible. It is extremely zygomorphic, and the
various species show different degrees of the phenomenon.
All show it in the stamens, but in H. Griffithii the fruit is
only very slightly zygomorphic, though as it ripens it
becomes more or less prostrate ; in H. sessile it is aniso-
lobous and erect, and in H. olivaceum much the same, while
in H. lichenoides its zygomorphism is extreme, correlated
perhaps with the fact that in this form the secondary shoots
are often comparatively prostrate throughout their life. In
Farmeria the morphological construction is similar to that
in H. lichenoides, and here we find the most extreme dorsi-
ventrality in the family, though the flower still curves
upwards when the spathe splits. The lower loculus of the
438 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
ovary is much displaced towards the lower side, a pro
that may be seen in an incipient stage in Mniopsis, &c,, anv^
its ovules are completely absent, while those of the upper
loculus are reduced in number and increased in size, and
the placenta is developed mainly at the distal end of the
ovary. In this form dorsiventrality appears at last in the
very embryo.
This remarkably interesting series thus seems to show
clearly that in the whole of this family dorsiventrality is
more or less strongly marked, and that its impress is first
shown in the vegetative organs, i.e.^ in the thallus and
secondary shoots, followed by the bracts, spathe, and flower,
while in the latter the stamens first and most commonly
exhibit it, followed by the stigmas, carpels, interior of the
ovary, fruit, and finally seed or embryo. Always the
dorsiventrality is most strongly marked in the vegetative
organs, and is often present in them to a slight or even great
degree (Lawia) without necessarily showing in the floral
organs, while when the latter are exceedingly zygomorphic
the vegetative organs are usually still more so.
To put the matter in other words, it would appear that the
dorsiventrality of the flowers, ivhich is the most important
morphological character in the classification of this order,
is a direct result of, or in direct correlation with, that of the
vegetative organs, being greater the greater the dorsiven-
trality of the latter.
This conclusion is confirmed by the argument from the
ecological relations of the flowers. We have seen above
that the dorsiventrality of the floral organs appears to be of
no value to them, but if anything a positive disadvantage,
and that as it becomes more and more marked the flower
becomes more and more self-fertilizing by aid of the wind
or by simple contact, while in Podostemon Barberi it is
cleistogamic. At the same time, too, the flowers stand erect,
an unusual feature in such zygomorphic forms .
It seems, then, not unreasonable to conclude that the
do)‘siventra,lity of the floral organs has Ijeen, so to speak.
OF CEYLON AND INDIA.
439
Upon them by that of the vegetati ve organs, or by
üûeir position upon the latter, without any reference to
advantages or disadvantages to he derived from it in the
performance of the functions of the floral organs themselves.
The only demand made upon them, so to speak, has been
that they should not cease to set seed ; this they have
continued to do with success, but in the more dorsiventral
flowers the chance of cross-fertilization is much diminished,
and even in some cases lost altogether.
To turn for a moment to the general conditions of life
described in the introduction, the total of these, acting upon
submerged organisms with creeping roots and secondary
shoots and with subaerial flowers and fruits, would seem
almost to necessitate more or less of dorsiventrality, and as
the vegetative organs are in more direct relation to these
conditions, the effect may be expected to show in them first
or only ; that it shows also in the floral organs is perhaps
an accident of the physiological nature of these plants or of
plants in general. Whether the effect on the floral organs
is direct, or through the vegetative, we cannot say. It
would seem perhaps most probable that the dorsiventrality
of the floAver has not only followed but been determined by
that of the vegetative organs and the bud-position; it never
reaches any great structural intensity in cases where the
dorsiventrality of the vegetative organs or of the position
of the developing flower is not also highly marked. It
apparently shows later in the evolution than the dorsi-
ventrality of the vegetative organs, but it seems gradually
to work through the series of floral organs, beginning with
the bracts and spathe and sho wing at last even in the interior
of the fruit and in the seed.
It should be especially noticed that the structural
dorsiventrality of the flowers is accompanied by two
other phenomena which are always regarded as of high
morphological and taxonomic importance, suppression of
parts (here shown in perianth, stamens, carpels, ovules)
440 WILLIS : MORPHOLOGY OF THE PODOSTEMACBÆ
and union or branching of stamens or receptacular prolon-
gation (according to the particular interpretation put upon
the forked andrœceum of most of the Eupodostemeæ).
We can scarcely suppose the floral organs to have led the
way in the dorsiventrality, because throughout the family
we find the vegetative organs to be the more dorsiventral,
and we have also seen that the floral organs appear to
derive no advantage from their zygomorphism ; further,
zygomorphism is almost unknown in floral organs that
are developed in erect positions on radially symmetrical
organs.
The general conclusion we have thus drawn is led up to by
two separate lines of argument, one from the morphological
construction of the organs in the whole series of Podostemaceæ,
the other from a consideration of the ecological features of
the flowers and fruits. If it be correct we ought to find
confirmative arguments from the general phenomena of
dorsiventrality in other plants. A general comparison of
these phenomena shows at once that dorsiventrality is much
more common in the vegetative organs than in the floral,
but that it is also frequent in the accessory organs of the
flower, less so in the essential organs, especially in the fruit or
seed. Again, within the same family, compare, for example,
the slight dorsiventrality of the vegetative and floral organs
of Potamogeton with the great dorsiventrality of those of
Zostera and others. There are many cases in which the
vegetative shoots are not themselves dorsiventral in structure,
but the zygomorphic flowers which they bear are developed
in horizontal positions, a feature which we have seen
to be one of importance in determining the degree of
dorsiventrality in the Podostemaceæ ; even in these cases it
is not to be forgotten that there is marked dorsiventrality
in the leaves.
One of the marked features in the flowers of the
Podostemaceæ is that the absent stamens are always those
of the upper side of the flower ; this again is an almost
universal feature in other dorsiventral flowers, e.g,, in the
OF CEYLON AND INDIA.
441
Tubifloræ, iu the Orchidaceæ, Liliaceæ, Musaceæ, &c. The
flower of Stellaria media is usually regarded as radial in
symmetry, but Burkill,'" by examination of a great number of
flowers, has shown that there is a distinct tendency for the
upper stamens to be most often absent ; here perhaps we
have the phenomenon in its incipient stage.
Again, in the Podostemaceæ the upper stigma is generally
smaller than the lower, and this is a very common pheno-
menon among zygomorphic flowers.
With regard to the immediate mechanism of the production
of dorsiventral structure, we as yet know but little.
Vöchting has shown,! in certain cases of slight floral
dorsiventrality, that rotation upon the clinostat during
development will entirely prevent the curvatures of the
organs concerned from taking place. Other phenomena have
been noticed elsewhere, which tend to show that gravity has
at least been an active factor in the production of dorsiven-
trality. In the peculiar thalli of Dicræa stylosa fucoides and
Griffithella Hookeriana Willisiana above described, the
thallus is usually drawn out most upon the lower side, while
it is symmetrical if it grow in an erect position.! Light too
may have something to do with the production of dorsiven-
tral structure, by retarding the growth of the illuminated
side, or in some more direct way. Further physiological and
experimental morphological work is required before we can
draw any definite conclusions in regard to the mechanism of
the production of these or other morphological phenomena.
In any case, the various possible agents have all a clear field
for the full exertion of their action in most of the Podoste-
maceæ, in consequence of the peculiar mode of life of these
plants.
It would seem then most probable that to a large extent
dorsiventrality in the floral organs is due, not to natural
* Variation in the number of stamens and carpels. Journ. Linn. Soc.,
XXXL, 1895.
t Ueber Zyg-omorphie Prints. Jahrb., XVII., 1886.
f And cf. Goebel on Stereum, Flora, 1902, vol. 9Ü, p. 471.
44:2 WILLIS : MORPHOLOGY OP THE PODOSTHMACEÆ
selection in reference to the performance of the functior.
the flowers, but to the direct action of other causes, and thai
directly or indirectly it is mainly the result of dorsiven-
trality of the vegetative organs and horizontal position in
development. If we accept this view, it renders much more
easy the task of explaining the stfuctural peculiarities of
many zygomorphic flowers, as we no longer need attempt to
explain every detail of structure in reference to the visits of
insects or to the advantages of cross-fertilization, and can
easily explain many such cases as those in the Labiatæ, where,
though (following the rule above indicated) the upper
stamen is absent, the working of the mechanism of the
flower is effected by having the stamens on the upper side,
and so on.
If we do not accept this explanation, we have to find
another to deal with the very remarkable facts. We can
hardly suppose that the process of evolution in the order has
been from the markedly dorsiventral to the less dorsi ventral
forms, for the former exhibit all the peculiar features of the
order together, and are perhaps the most aberrant types
among the flowering plants. If we do, however, accept it,
we are driven then to suppose that the floral organs have
become radial and entomophilous, and that the vegetative
organs have followed by losing their dorsiventrality, though
we know that the more dorsiventral they are the better they
are suited to their peculiar mode of life, and the less
competition the species meets with. This, then, would be a
case of characters of great importance being forced upon the
vegetative organs in spite of their disadvantage to those
organs, or at least non-advantage. Probability, analogy, and
the general evidence are all in favour of the former view,
that it is the floral organs whose advantages have been
disregarded in consequence of the imperative claims of the
vegetative organs.
If we suppose the dorsiventrality of the flowers to have
been the subject of natural selection, independently of th e
vegetative organs, we have to explain the very remarkable
OF CEYLON AND INDIA.
44^
fact that its degree seems to depend upon that of tlie vege-
tative organs, and that it is never found without the latter,
to say nothing of such difficulties as the erect position of
most of the flowers and the apparent disadvantage of their
dorsiventrality to them.
It is evident that if this general contention be accepted as
probable it opens up many new points of view, and many
lines of research. There is as yet no evidence to show how
the reaction affects the actual sporangia, the gametophyte, or
the essential sexual generative cells ; the question requires
investigation, in view of the fact that the lower groups of
plants, from which ihe higher are supposed derived, are
so largely dorsi ventral in structure. The general evidence
above seems to indicate that in the higher plants the dorsiven-
trality is derivative, and that radial structure has in general
preceded it.
If, now, we accept the dorsiventrality of the flower as in
great degree the consequence of that of the vegetative organs
— let the mechanism or correlation be what it may — we are
led to some important deductions which once more raise
several questions which have of late been formulated from
other considerations, and which must be set at rest.
If one character — or two, for suppression seems forced
upon the flower of the Podostemaceæ as well as zygomor-
phism — of great importance in morphology and taxonomy
may thus be forced upon an organ without reference to its
advantage or disadvantage in the performance of the func-
tions of that organ, it seems only likely that other char-
acters may have been forced upon the same or other organs
in similar or other ways, and consequently that the study of
adaptation must enter upon a new phase, or rather, must be
accompanied and checked by the study of experimental and
comparative morphology, and of variation and correlation,
if we are to expect further valuable results. The study of
correlation has already yielded many interesting facts, but
in ontogeny rather than in phylogeny ; the case of the
dorsiventrality of the flowers of the Podostemaceæ seems at
(59)
444 WILLIS : MORPHOLOGY OF THE PODOSTBMACEÆ
last to provide a good instance of fixed or phylogenetic
correlation, to support the few yet adduced.
It is also evident that the familiar statement that the
characters of an organ or organism are partly hereditary
and partly adaptive does not cover the whole truth, and
that we must allow for the presence of what we may perhaps
term correlative or induced characters, characters due to
correlation or to the parallel action of similar causes, and
not necessarily advantageous, though fixed in the heredity.
The objection is often brought against the theory of
natural selection that many specific characters — and less
often that generic characters — are useless to their possessors.
Hitherto, however, so far as I am aware, few cases have been
adduced in which it has not been possible to reply that
further investigation may prove the usefulness of these
characters, or that they may have been derived from useful
characters in ancestral forms. In the few cases not to be
explained in these ways, correlation has been adduced."’ In
the Podostemaceæ wo have perhaps the best instance of some
specific and still more generic and tribal characters,
apparently due entirely or almost entirely to correlation or
to direct action of external causes, and not to natural
selection. Natural selection, however, may well have acted
largely on the vegetative organs, which are most directly
concerned with the external conditions of life.
We are perhaps, at least in Botany, too much in the habit
of regarding evolution from an extremely analytical point of
view, taking the characters of an organism singly, and more
or less unconsciously assuming that the evolution has
affected one at a time, as represented by a diagram of
dichotomous branching in one plane, such as that given by
Darwinf to illustrate the possible formation of new species,
and thus implying that a change in any organ must neces-
sarily be accompanied by some advantage to that organ or
* Cf. Darwin, Origin of Species, 6th ed., p. 116 ; Wallace, Darwinism,
ch. VI,
t Origin of Species, 6th ed., p. 90,
OF CEYLON AND INDIA.
445
to the organism of which it forms a part. We ought rather to
consider the whole organism or species together, remember-
ing that there may be for it a kind of general balance sheet
of profit and loss,* in which it is not necessary even for
success and expansion that there should be a profit under
every item, but that the species will lose or gain according
to the balance on the total of all the items, some of which
may in themselves be very disadvantageous. In the Podos-
temaceæ we have a striking case in point. Once in the
rushing water, the general conditions of life and variation,
in conjunction with the hereditary peculiarities of the
family, have produced, so far as we can judge, a certain
general type of vegetative structure and habit suited to the
environment, while the floral system has been forced to
accommodate itself to the changes of the vegetative, the only
demand made upon it being that it should continue to set
plenty of seed and try for cross-fertilization as much as
possible under the circumstances. The disadvantages of the
floral construction forced upon these plants have apparently
been made up for by the advantages of the vegetative.
Another interesting point brought out by this method of
looking at the facts is the insecurity of the basis on which
our conceptions of species (at least the large “ Linnean ”
species), genera, and still more the higher groups rest,
and the great need for a more exact and quantitative foun-
dation in our science. Supposing in the case in question
that the dorsiventrality of the environment and heredity
were to act only on the vegetative organs, without any
correlative or induced action upon the floral organs, Ave
might then have to include almost all the Podostemaceæ in
perhaps three or four genera at most, as we should only
have a very few marked distinguishing points, e.g.^ the
presence or absence of perianth, the distinction betAveen
shoot- and root-thalli, betAveen thread-like and ribbon-like
* Cf. Bateson, Materials for the Study of Variation, p. 12 ; MacLeod in
Bot. Jaarb, V., or abstract in Willis, Manual and Dictionary, vol. I.,
p. 16.
446 WILLIS : MORPHOLOGY OF THE PODOSTEMACBÆ
thalli, and so on. Most of the present genera and tribes
would be united with others, or perhaps the tribes might
become genera, the genera species, though in nearly all
their essential features, i.e., in those related to the environ-
ment and to the mode of life, the plants would be as different
as ever. This insecurity must remain until the study of
experimental morphology with variation and correlation
shall provide us with a more reliable quantitative test of
generic and other differences. On the whole, perhaps, our
specific and generic grouping will be found ultimately
supported by the facts of evolution so far as they may be
discovered, but it is well at times to remember the insecurity
of our position. It is partly in view of this insecurity of
the larger grouping, that though in the classification of the
Indian Podostemaceæ in the preceding paper I have used
the large Linnean species, I have used proportionately small
genera, in which vegetative characters have been possible of
use in addition to the floral.
The next question that may be raised is whether many of
our groups, such as the Linnean species, and still more
the genera, tribes, sub-orders, orders, and larger groups are
or are not polyphyletic. The question is already being
raised by the fact that so many of the largest groups, e.g.^ the
Fungi, the Gymnosperms, the Sympetalæ are proving to be
polyphyletic, so far as the evidence of palæobotany and
morphology enables us to judge, but the considerations
above given raise it from a different standpoint, and enable
us to see one way in which the phenomenon may have
arisen. We can easily imagine a case like the following.
Let A be a species of Podostemaceæ, Avith a marked but not
extreme dorsiventrality of the vegetative organs, and let it
give rise by gradual evolution to three species, K, L, M,
of which K and M are more, L less, dorsiventral than itself in
the vegetative organs, while in all other respects K is more
nearly allied to L than to M. Now let these go on to evolve
further species, X, Y, Z, respectively, the dorsiventrality
still increasing in the case of the first and last, but X and Z
OF CBYLOIS^ AND INDIA.
447
remaining nearly akin to Y in every other respect. The
increasing dorsiventrality will at least re-act on the floral
organs — let the mechanism be what it may — and the latter
will show increased dorsiventrality. Now the phenomena
of this increasing dorsiventrality seem, as we have seen, to
follow fairly deflnite laws, and hence it is by .no means
unlikely that the same phenomena will show in X and Z,
while they do not show in Y, which has not increased its
vegetative dorsiventrality. We shall thus get a polyphyletic
genus formed, with two species X and Z, which may be each
as nearly akin to Y as to one another, though Y is not
included in their genus. The genus will be polyphyletic in
that there is no common ancestor which possesses the generic
characters.
There is no necessity, perhaps, for the case to be so simple
as that just sketched. It is quite possible that species not
at all closely allied may take the same generic step, so to
speak, which brings them close together in our schemes of
classiflcation, though their immediate ancestors might have
been far apart by the criteria used.
The same reasoning will of course apply to the case of a
number of forms taking a similar tribal or sub-ordinal step
when the conditions occur, and in fact the larger the group
and the fewer the characters on which it is based, the greater
the likelihood of its being polyphyletic." It is by no means
unlikely that some of the many cases of genera where the
different species seem to have each of them some of the
generic characters of surrounding genera, while all have
their own generic character, may be due to some such process
of evolution as that just sketched. Thus in Hydrobryum it
is quite possible that the smooth-fruited H. sessile may belong
to a different phylum from the other species with ribbed
fruits, or in Farmeria, that the two species may be derived
from different phyla, though of course the facts may be
explained on other hypotheses. Just as an idea or a discovery
An old rule in a new and more easily-fitting dress.
448 WILLIS ; MORPHOLOGY OP THE PODOSTEMACEÆ
is very often made approximately at the same time by more
than one mind acting on somewhat similar materials in
similar ways, so we may imagine the same generic or ordinal
step taken by allied or similar species at about the same
time or in relation to similar causes. This of course is only
a supposition, but the question as to whether many genera
and other groups are or are not polyphyletic is raised on
sufficient evidence to require that it shall be investigated
and set at rest, as well as the further question, whether, if
any genus or other group is found to be polyphyletic, it is
to be retained as an artificial group or divided into others,
even though there may be the very closest similarity among
its members.
It is evident that if the phenomena here suggested are at
all common, the process of evolution may in these cases have
been much swifter than we are usually inclined to believe,
many forms perhaps having undergone parallel develop-
ments, so that the evolution may be likened to the climbing
of steps by several individuals at once.
To follow up the same line of thought, it is clear that it
may considerably alter some of the prevailing views upon
geographical distribution, for if the generic or ordinal step
may be taken separately by similar or allied species, these
species may be already separated by large distances, and
the genus or order may thus conceivably arise in two or
more phyla at different places. It is no longer absolutely
necessary to suppose, to take the particular case in hand,
that the extremely zygomorphic-flowered Eupodostemeæ
must have spread from a common centre occupied by one
original ancestral species with the sub-ordinal characters ;
there may have been, in India and in the other regions
where this sub-order is found, several species which under
the influence of similar causes may have taken the steps, so
to speak, which transformed them to representatives of this
sub-order. All that is necessary is the final stage in the
zygomorphism of the andrœceum, reducing it to one single
or double stamen on the lower side of the flower.
OF CEYLON AND INDIA.
449
Indications are accumulating from many quarters, from
palæobotany, from morphology, from the study of variation,
and from other considerations, to show that the question of
the possible polyphyletic origin of many groups hitherto
supposed monophyletic, whether large species, genera,
orders, or higher groups, is one which will soon demand
much attention. Should this phenomenon prove to be at
all common, it will of course explain many problems which
have hitherto been great difficulties to students of evolution
or taxonomy.
All these suggestions of course contain a large element
of speculation, but if we admit, and the evidence in its
favour seems fairly strong, the first hypothesis, that the
dorsiventrality (including suppression of organs, &c.) of
the flowers, fruits, and seeds of the Podostemaceæ has been
very largely determined by causes having nothing to do
with their functions, while at the same time we regard these
features of the floral structure as our great criteria in
classification, all the questions above raised follow inevi-
tably, and once raised on good evidence they must be settled
one way or another before we can feel secure as to the
stabilit}^ of the foundation upon which so imposing a structure
has been raised in the sciences of biology. If we do not
admit the hypothesis, we must find another which will as
well explain the facts, themselves so striking that they
demand explanation.
It is clear that we require a detailed investigation of
most of those comparatively constant structural features,
especially in the floral organs, upon which we base our
classifications, in order to discover among other things,
whether, as seems to be the case in dorsiventrality, there
is any correlation between them and any of the features
of the vegetative system, or any other factors which may
have been more or less causal ; any such correlation once
traced or even imagined, the phenomenon should be the
subject of physiological and experimental morphological
investigation, as well as of variational studies.
450 WILLIS : MORPHOLOGY OF THE PODOSTEMACBÆ
The Geographical Distribution of the Indian
Podostemaceæ .
Owing to their being confined to inonntain streams, and to
their somewhat imperfect mechanism for seed-dispersal, the
Podostemaceæ are very local in their distribution, and as has
already been several times mentioned, each river or group of
rivers or even each section of each large river often possesses
its own local forms or species of each genus represented in it.
With a complete knowledge of the forms in all the mountain
streams of the tropics, it should be possible therefore to
arrive at conclusions of great importance as to geographical
distribution within these regions, but at present only the
beginnings of such knowledge are at our command.
The only method of distribution which appears to be of
any avail to carry these plants from one river to another is
the adhesion of the sticky seeds to the wet feet of wading
birds ; these may often be seen in the dry season walking
over the thalli, and the seeds must cling to their feet. It is
therefore unlikely that any carriage of the seeds over more
than a few hundred miles ever takes place, and in general
the distances travelled will be much less. As in India and
Ceylon the flowering takes place in the north-east monsoon,
and as it is easier, in all probability, for a plant to suit itself
to a longer than to a shorter vegetative period, it would seem
not unlikely that the general migration may be southwards
rather than northwards ; but there are numerous other
factors to be taken into consideration, and it is not safe to
say more than that perhaps this is the tendency at the present
time.
The order is practically confined to the tropical regions
of the globe, and only’ passes into the warm temperate zones
in a few districts. The local distribution of the forms
is of considerable interest, and helps to throw light upon
their evolution and adaptation. So far as India and Ceylon
are concerned, we find that in general the larger forms
OF CEYLON AND INDIA.
451
live in the lower levels of the country, in the larger streams,
and in the more southern districts, Le., in the districts
with the best distributed rainfall and the longest vegetative
season. This may be best shown by a few examples.
Thus in Ceylon the species that occur at the highest
levels are Hydrobryum lichenoides and H. olivaceum^
while those that descend lowest are the large Dicræas. In
the Nilgiris H. olivaceum is found at least as high as any,
accompanied by the most prostrate of the Indian Dicræas, D,
dichotoma, while the large species of the Malabar region,
Tristicha ramosissima, Podostemon Barberi, Dicræa stylosa,
&c., occur at low levels. Or again, to consider the size of the
streams, whether in Ceylon, South India, the Bombay Ghats,
or Assam, the only species found as yet in really small
streams (often in mere becks of a yard or two in width) are
Hydrobryums, Lawias, and Farmerias, with rare exceptions.
Again, to deal with the latitudinal range, the species which
go furthest into the more northern regions of short vegeta-
tive season are also the Lawias and Hydrobryums, while the
larger genera are confined to the more southern regions.
Lastly, consider the immediate local habitat of the species
growing at one place, e.g., at Hakinda. As has been fully
described under the individual forms, each form has in
general its own peculiarity of habitat, and is but little mixed
with the others, except at places where the conditions vary
very much in a small area, for instance where an eddy joins
a swift current in the main stream. As will be easily seen by
reference to the details of habitat given under each species,
the larger forms live in the less rapid and violent water,
which of course is also the less liable to rapid shallowing.
It is thus evident that the dwarf forms affect, and are
apparently best suited to, swift streams and rapid water which
may easily or quickly become shallow. The question then
arises, whether their adaptation is to the swiftness or to the
shallowness, or to both. Probably, as has already been
pointed out, the latter condition is the more important, as
even the large forms can stand much swifter water than what
(60)
452 WILLIS ; MOKPHOLOdY OP THE PODOSTEMACEÆ
they usually affect without difficulty, and some of the dwarf
forms, e.^., Hydrobryum olivaceum, only become completely
dwarf as the water-level is falling, and have tall shoots
at the season of deeper and swifter water. Some of the largest
American forms inhabit very rapid water.
With regard to the local distribution of the different genera,
it may suffice to note that the richest region is the south-
western corner of India, together with Ceylon ; here there
occur three endemic genera, Lawia in the Tristicheæ, and
Griffithella and Farmeria in the Eupodostemeæ, while there
are endemic species in Tristicha, Podostemon, Dicræa, and
Hydrobryum (and probably Willisia). Only the three last
occur in the eastern peninsula. It will be noticed at once
that the distribution of the Podostemaceæ in Ceylon and
South-West India is just like that of the so-called Malayan
element of the Ceylon and Malabar flora, rich in the south and
fading out to the north. Only one species has as yet been
recorded from the Malayan region (30), but this shows
affinities to the Abyssinian Sphærothylax. Tristicha and
Dicræa have their nearest relatives in Madagascar, while
Hydrobryum, the most widely distributed of the Indian
genera, is related to Sphærothylax. It would seem, therefore,
not impossible that the order reached India and Malaya
(including Burma and Assam) from the south-Avest, but if
so, there must almost certainly have been, in the land forming
the passage betAveen India and Madagascar,’^' some hilly or
mountainous country, for even Tristicha and Dicræa require
fairly rapid water Avith a rocky bottom.
Th® Origin of the Podostemaoeæ.
Tliis is a someAvhat difficult question, which must be left
for detailed consideration in connection with the anatomical
inAæstigations yet to be completed. The embryonic develop-
ment and characters are so typically dicotyledonous, and
Cf. Willis and G-ardiner. The Botany of the Maldives. Ann, Perad., !..
190L pp. 5, 141, HI
OF CEYLON AND INDIA.
453
the other characters of the flowers and fruit, &c., are so
simple, that one must regard the order as representing an
old phylum of true Dicotyledons, The next question that
arises is whether it has descended from aquatic or terrestrial
ancestors. On the whole, considering the terrestrial type
of both flower and fruit, the latter seems most probable.
A possible origin for the order seems to be from plants
already growing on the banks of mountain rivers, with
creeping adventitious roots upon which secondary shoots
were regularly developed. We can imagine these plants
“ taking to the water ” by means of these shoots, which in
the intermediate period of transition would have to be more
or less amphibious. Possibly the anatomical difference
between the primary axis and the secondaries may indicate
that the latter became adapted to the new existence sooner
than the former. . We do not yet, however, so far as I am
aware, know to what extent primary and secondary axes
differ in other plants. The anatomy of the primary axis in
the few Podostemaceæ yet examined is moi’e like that of
other water plants than that of the secondaries, which
perhaps indicates the origin of the order from some form
akin to the Nymphæaceæ or other ancient group of water
plants, far back in the evolution of the Dicotyledons.
Tl%e Systematica Pesitiaii tiie Pedastemaeeæ.
This also is a question for further discussion when fur-
ther evidence shall be available. The family is very isolated,
like so many other groups of submerged water plants. In
our present ignorance of the phyla of the evolution of the
higher plants, the most we can hope to do is to trace some
line of apparent descent near to which we may place the
order. I have already indicated"' that I regard the Sym-
petalæ, as do many other authors, as polyphyletic ; one of
the lines of their phytogeny seems to me to run through the
neighbourhood of the Englerian cohorts Rosales, Myrtifloræ,
* Manual and Dictionary, vol. I., p. 68.
454 WILLIS : MORPHOLOGY OF THE PODOSTBMACEÆ
and Contortæ, connecting to the Tnbifloræ. Now there is
little doubt that the Podostemaceæ have some affinities to
the Saxifragaceæ, as pointed out by Warming (43), and to
the Lentibulariaceæ, as pointed out by Hooker (18), while
long ago Gardner (12) pointed out what is possibly their
nearest relationship, viz., to the Nepenthaceæ and the allied
orders, such as Sarraceniaceæ. These are tropical orders,
and the ancestors of the Podostemaceæ were probably also
tropical. On the whole, therefore, I think that the order
may well be placed near the cohorts Rosales and Sarrace-
niales, but the whole question requires much further
investigation.
The Qenea^aS E^elistiosi ef the Pedostemaeeæ.
To sum up in a few words the general bearing of the above
discussion upon the question of the evolution of these plants
in reference to the general conditions of life which were
considered in the introduction, we may perhaps regard it as
not improbable that the order is descended from tropical
Dicotyledonous plants rather far back in the evolution of
the Dicotyledons. These plants may have been aquatic or
may have been terrestrial ; they probably had creeping
adventitious roots bearing secondary shoots, and capable, like
most such roots, of rejuvenescence after any injury to the
apex ; their flowers were probably erect, regular, with simple
perianth, insect-fertilized ; their seeds small, ripening in
the dry season of the year. Gradually we can picture these
plants taking to life in moving water with a rocky bottom,
to which their roots would fasten them, and later, as the
water strain became greater, developing haptera (perhaps by
some modification of roots) to resist this strain. The roots,
thus becoming more exposed to the action of light and
contact, as well as of gravity, would probably, like the
creeping roots of orchids, &c., tend, under the influence of
one or more of these causes, to become more or less flattened
and chlorophyll-bearing. Accompanying this expansion of
the root and its assumption of assimilatory functions there
OP CPYLON AND JNDIA.
455
would appear to have been usually a dwarfing of the secondary
shoots, ultimately going so far as to reduce their axes almost
to nothing, as in the Dicræas and the Farmerias. As this
dwarfing progressed, the plants would become more and more
suited to life in shallower (and in general also swifter) water,
and would spread into regions of shorter vegetative season,
into the higher regions of the hills, and into the smaller
streams, as well as into the swifter w^ater of the larger
streams. In this way they would escape more than ever from
competition, and perhaps be enabled to evolve in greater
variety of form than might otherwise be the case.
The dwarfing at the same time would seem to have been
accompanied by an ever-increasing dorsiventrality of the
vegetative structure, the roots becoming more and more
flattened and thalloid, and taking over more and more the
functions of assimilation. This dorsiventrality would seem
to have ultimately re-acted on the floral organs, which became
more and more zygomorphic, Avith suppression and other
important phenom.ena accompanying the zygomorphism . At
the same time the flowers appear to have become more and
more anemophilous and self-fertilized. Most, if not all, of
the changes in floral morphology on Avhich the groupings of
the order are largely based would seem thus to be correla-
tive or induced, and not to have been the subject of natural
selectiou, unless in a very minor degree, a conclusion which
opens up interesting general considerations.
In conclusion, I wish to thank, in addition to the friends
mentioned at page 182 of the preceding paper, my wife,
without whose help this paper Avould never have reached
its present state of comparative completeness, Professors
Warming and Goebel, and Dr. Treub, to whom I am indebted
for references and criticisms.
Peradeniya, July 15, 1902.
456 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
Explanation of Plates IV -XXX VIll.
Note, — In consequence of its having been decided, aHer some of the later
plates had already been printed, to omit certain anatomical details, it has
been found necessary to omit Plates XXIII., XXVII., and also the last two
Plates (already mentioned in the preceding paper; XXXIX. and XL. The
total number of plates published with this paper is thus 33.
All drawings and photographs are the work of the author, with the
exception of fig. 2. PI. XXX.. for which he is indebted to Mr. A. de Alwis.
It may be noted that the photogravures, being prepared directly from the
negatives, show much detail, and will bear a magnification of two or three
diameters under a lens.
Owing to the great local difiierences in these plants the source of the
material is quoted in most cases.
Plate IV. — General view of part of Ilakinda rapids on the Mahaweli-
ganga near Peradeniya (the background is part of the forest reserve of the
Experiment Station), showing Podostemaceæ exposed upon the rocks in the
dry season.
The island rock in the left-hand bottom corner is covered with fruiting
Dicræa elongata ; the rock to the extreme left, in the sheltered bay, is
covered with Podostemon subulatus. To the right of it, in the less rapid
part of the main current, is Dicræa elongata. The rocks at the little fall
(which is about 3 feet high) are covered with Hydrobryum olivaceuna,
while the parts between the fall and the Dicræa bear both this species and
Lawia zeylanica, as also does the central island.
Plate V., Tristicha ramosissima ; herbarium specimen prepared by
Mr. C. A. Barber by floating out under water (material from S. Kanara,
No. 2,518), X 4/7. This gives a very clear idea of the general habit in the
vegetative condition ; the creeping root-thallus runs through the middle
of the sheet, giving off secondary shoots (often paired) with hapterous
bases, and bearing branches of two kinds, long and short, the latter
(ramuli) with filamentous leaves.
Plate VI. — The same. The left-hand specimen shows the branching
clearly ; at each node two branches, long and short, the latter below the
former. The right-hand specimen shows the floral and fruiting shoots,
and the way in which the ramuli lose their leaves at this season. The
floral shoots have a few leaves at their bases, and a couple of ramuli, and
stand erect, emerging from the water in which the main shoot is drifting.
Plate VII., Tristicha ramosissima. — Fig. 1 Portion of thallus
(th ) attached to rock, showing hapterous discs or feet (/e of the thallus, x 6. 9 A. 2-.
Parkhiiana, a specimen showing the branching and the rosettes, x 8/7.
Plate XII., Lawia zeylanica Gardneriana. — A museum specimen
preserved in alcohol, X 5/6. Young plants in vegetative condition,
gathered in August or September at Hakinda, showing the mode of
branching, the general form, and the flabelliform apices. Some of the
large lobes show division by small hedges of leaves (represented by white
lines) forced upwards by the lobes against one another.
Plate XIII., Lawia zeylanica. — 1 Apices of var. konkanica, in fruit,
X IV 2 The same in section to show the shallow cupule. 3 Cupule and
open flower of var. malaharica (C. A, Barber, No. 2,150), showing the
stamens closely pressed against the stigmas at anthesis, X 5g. 4 Trans-
verse section of vascular strand in pedicel of flower, x 600, showing
parenchymatous cortex (y.c.), lignified tissue (Ug.), and medullary cavity
(m.c.). 5 Oupule of L. z. Gardneriana^ x 6, before emergence of flower.
6 Flower of L. z. lionkanica (Igatpuri) emerging under water from
shallow open cupule, x 5. 7 Fruit of the same, X 5. 8 L. z. Gardneriana.
longitudinal section of cupule (cuf) and included flower, showing leaves
(/^.), anthers (anth.). stigmas Gtig.). ovary (g.)^ and perianth (p r.), X 30.
OF OFYLON AND INDIA.
459
9 Transverse section of the same, diagrammatic, X 20 ; r.h.^ the vascular
bundles. 10 Transverse section of flower of L. z. konkanica (Kasara),
X 4.5, showing perianth, stamens, ovary, placenta, vascular bundles, &c.
Plate XIV,, Podostemon subulatus.— 1 (left-hand) Stone with
fruiting stalks, X 2/3. Most of the fruits have lost one valve. The
secondary shoots are more or less concealed in a mass of dead shrivelled
leaves (c/. PI. XV.). Dicræa elongata. — ^2 (small upper left-hand stone)
Fruiting specimen, in which the creeping thalli are bearing fruit, for
comparison with fig. 3 and PI. XIX., X 2/3. D. dichotoma. — 3 (upper
right-hand) Fruiting specimen, X 2/3, showing creeping thallus with
distichous fruiting shoots ; the free drifting ends of the thalli mostly
absent. D. stylosa kanarensis. — 4 Fruiting specimen, x 2/3, showing
long pedicels inmost cases, but with some short ones. The broad alga-like
thallus can be seen in the upper left-hand part.
Plate XV., Podostemon subulatus Mavœliæ, x 4/5.— Plant
gathered at Hakinda in August, in full vegetative growth, showing the
long distichous leaves and tufted habit. To the left are isolated secondary
shoots, united at the base by the thread-like creeping thallus.
Plate XVI. — The same. 1 Seedling, x 5^, showing cotyledons {cot.') and
first two leaves. 2 A larger seedling, x 5^, with beginning of thallus (th.).
3 An older seedling, with thallus bent into line with axis, x 5^, showing-
branching of thallus and first endogenous secondary shoot (.9.6*.). 4 Apex
of a thallus, X 16, to show origin of secondary .shoot. 5 An apex, from
above. 6 Portion of a young plant, x 4, showing the way in which the
thallus branches right and left alternately, and the origin of the shoots
<^cf. fig. of Bostrychia Moritziana in Goebel’s Organographie, p. 32).
7 Transverse section of thallus, X 20, showing outline, vascular strand (v.h.),
and the foot formed by the root-hairs (/.). 8 Transverse section of secon-
dary shoot through leaves at apex, x 10. Plant in vegetative condition,
gathered 1-9-08 ; the concavity of the shoot is towards the upper side, and
the leaf-sheaths are thicker on that side ; v.h., the vascular strands.
9 Transverse section of pedicel of flower, X 150, showing epidermis
more columnar on upper side, parenchymatous cortex {p.c.), and vascular
bundle (v.&.). 10 Flower in open spathe x 6 ; l.l., the leaf in
whose lower axil it arose ; atcl.., the staminodes or perianth segments at
sides of andrœceum. 11 A fruiting secondary shoot in dry season, nat.
size. 12 Fruit, X 6, from side that was upper in bud, showing back of
larger lobe, with its three decurrent ribs. 13 The same from the side,
showing the way in which the smaller valve breaks away.
Plate XVII., Podostemon Barberi.— 1 Part of a plant, nat. size,
slightly diagrammatic, showing thallus (th^^ secondary shoots (s.s.), and
flowers. 2 Flowering shoot, X I5, the flower still enclosed in the spathe
{spa^. 3 Transverse section of thallus, x 20, showing vascular bundle
(y.y.) and mass of rhizoids (/.). 4 Transverse section of thallus, x 90.
5 Flower in spathe, the front of which is removed, X 6. 6 Transverse
section of the ovary, X 60, showing six-ribbed outline, the dehiscence ribs
(deh.'), the vascular bundles (r.&.), the inner epidermis (i.e.), and the
deciduous outer part of the pericarp (deoid.'). 7 Part of dehiscence rib,
X 150, to show method of splitting ; i.c., the inner epidermis ; deh., the
(61)
460 WILLIS : MORPHOLOGY OF THE PODOSTEMACEÆ
splitting layer. 8 Transverse section of the inner epidermis of the fruit,
to show construction of the wall, x 320. 9 The fruit, in transverse
section, x 60, for comparison with fig. 6 ; the outer tissues have fallen
away, and the rest have become woody.
Plate XVIII., Dicræa elongata. — 1 Seedling, x 6, showing com-
mencement of thailus (^^.). 2 Larger seedling, X 6, showing development
of first free drifting thallus (dr. th.') from the creeping thallus. 3 Base
of seedling, X 40, showing cotyledons (cot.')^ endogenous thallus and
exogenous hapteron (h.'). 4 Seedling, x with drifting thallus form-
ing at right angles to creeping thallus ; the primary axis. 5 Apex
of drifting thallus, X 5, with root-cap and origin of secondary shoots.
6 Transverse section of drifting thallus, X 20, showing parenchymatous
cortex (p.c.) and vascular bundle (v.h.'). 7 Cortex of drifting thailus in
transverse section (taken about middle of length, 1-9-98), X 80, show-
ing commencement of cell division. 8 Transverse section of vascular
strand, near apex of drifting thallus, gathered 1-9-98, X 320 ; X, primary
xylem ; col.^ collenchyma. 9 A small part of the same strand, lower-
down, X 320, showing commencement of formation of secondary tissue. 10
The same, older, x 320, with formation of secondary xylem nearly com-
plete. 11 Transverse section through thallus at exit of secondary shoot,
X 16. F.&., the vascular strand with branch to shoot ; col., the collen-
chymatous layer ; d.c., the part of the cortex which finally shrivels ; v.a.,
the growing point of the shoot ; 1., its leaves. 12 Transverse section of
ovary, X 60, for comparison with Podostemon ; i.e., the iuner epidermis ;
■o.h., the vascular bundles.
Plate XIX., Dicræa elongata, x 3/8. — On the left fruiting specimens
from Hakinda, gathered in dry season. On the right a plant in water,
gathered in September in vegetative condition, showing the creeping
thalli at base, with long drifting thalli, along whose edges are the little
tufts of leaves representing the secondary shoots.
Plate XX., Dicræa elongata. — 1 Piece of thallus, x 4, showing re-
juvenescence from the base of an old portion re-submerged ; th., creeping,
dr.th., drifting thallus. 2 Fruit, X 6, for comparison with that of
Podostemon (PI. XVI.), to show isolobous development and mode of dehi-
scence. D. dichotoma. — 3 A small piece of thallus and secondary shoots,
to show zigzag form often assumed. 4 Transverse section of thallus in
flowering season, x 20, to show deciduous (dec.^ and persistent (pers.) parts
of cortex ; v.h., the vascular strand. 5 Portion of flowering thallus,
showing projections under the floral shoots, X 1^. 6 Secondary shoot,
X 2, in leafy condition, just beginning to form a flower. 7 Later stage,
with spathe (s^a.} visible between upper pair of leaves with enlarged
sheathing bases. 8 Later stage, with leaf tips falling away. 9 Flowering
shoot, with bracts at base whose leafy tips have completely fallen ; spathe
open, flower fully expanded. 10 Fruit, X 6. Dicræa Wallichii.— 11 A
portion of a plant from herbarium material, X 3, showing form of growing
apices, and development of the secondary shoots, with vascular bundles
(v.h.'), the latter thickened where they lead to shoots, about to bear flowers.
12 A small portion of the base of a plant (Cherra), X 2, showing vascular-
bundles before the beginning of development of flowers. 13 The same
after development of the flowers (these not shown;. 14 Base of a plant
OF CEYLON AND INDIA.
461
from below (Weddell’s D. pterophylla, spirit material collected by
Hooker), nat. size, to show the stout foot with ascending thallus. 15 Open
flower, with spathe and bracts, X 2.
Plate XXI., DicræaWalliohii Khasiana, x 4/7.— On the left are two
stones from below water, Cherra, Dec., 1901, showing the closely appressed
broad alga-like thallns with marginal buds ; on the right three stones from
above water at same place, showing the fimbriate fruiting thalli, the inter-
mediate tissues having fallen away. The various single plants occupying
the centre of the plate show all stages in the process ; all are from Cherra,
The three small specimens in the left-hand upper corner are the originals of
Weddell’s D, pterophylla ; the larger lobed specimen below them opposite
the space between the rocks is an original specimen of G-riffith’s D.
Wallichii ; the rest were collected by myself in Dec., 1901.
Plate XXII., Dicræa stylosa. — 1 Apex of form laoiniata^ x 5, with root-
cap (r.c.), secondary shoots (.?.6'.),and vascular bundles 2 Seedling
of form fuGoides, X 5; liyp., the hypocotyl, with cotyledons (^cot.'), young
thallus and secondary shoots (.?.^.). 3 A young plant ofßicoides, nat.
size, showing general habit of the form that is only attached by a basal
foot, 4 Vertical section of foot of same, X 2. 5 Longitudinal section apex
of thallus of fucoides, diagrammatic, showing root-cap (r.c.) and vascular
strand (^’.&.), X 80. 6 The same apex, from above, with secondary shoots
[s.s.'). 7 Transverse section of thallus, x 10, to show general form when
young, and central vascular strand. 8 Transverse section of same thallus,
X 150, showing epidermis (ep.'), parenchymatous cortex {p.c.'), and
vascular strand with its two groups. 9 An older thallus, X 80, to the
right of the vascular strand (v.h.'), to show the way in which the cortex
grows in thickness and the thallus curves upwards at the side.
.. Plate XXIII. is omitted, see above.
Plate XXIV., Dicræa stylosa. — 1 Floral shoots of var./wc'ci(^6?À*, x 4,
one with open, one with closed, spathe (^S2)a.'). 2 Open fiower of var.
Bourdillonii, with vascular bundle leading to it (t'.5.), X 2^. 3 Portion of
plant of var. Uanarensis^ X 3, showing the acuminate form of the ripe
bracts, the irregular distribution of the fioriferous shoots, and the thick-
ening of the tissues leading to them (it?.). Dicræa minor. — 4 Plant or
part of plant, from spirit material collected by Hooker, X IJ. 5 A secondary
shoot ready to flower, X 2L 6 A shoot with ripe fruit, from G-ritfith’s dry
material, X 3. G-riffithella Hookeriana. — 7 Apex of a thallus (C. A.
Barber, 2,149), X 4, showing development of leafy and floral shoots. 8
Floral shoot of same form, X 6, showing bracts with leafy tips still
persistent. The flower is ready for fertilization ; it has become twisted
round in emerging from the spathe std.^ the staminode or perianth
segment on one side of fertile stamens. 9 Transverse section of ripening
ovary, x 20, showing absence of vascular bundles in the wall ; pl.^ the
placenta, with its ovules (ot'.) ; /., the thick-walled cells forming inner
layer of pericarp. 10 Part of pericarp enlarged, X 320, showing dehiscence
layer {delid)^ inner epidermis (7.e.), sclerenchyma {scl.')^ and thick-walled
cells (/.).
Plate XXV., Griflithella Hookeriana Willisiana, x 3/7 (C. A.
Barber. Nos. 2,515, 2.519). — 1 (large left-hand rock. 2.519) A single
462 WILLIS: MORPHOLOGY OF THE PODOSTEMACBÆ
subn erg-ed plant, in vegetative condition, showing the mode of growth
and branching, and more or less prostrate leafy secondary shoots on
margins of thallus. 2 (uppermost) Small plants (2,515) in ripe fruit,
more or less shrivelled, but showing much variety of form, most often
more or less discoid and prostrate. 3 (upper right hand) The same, but
with stalked and scalloped forms. 4 (lower right-hand) Creeping form
(2,519), with half -ripe fruits on the secondary shoots.
Plate XXVj., Griffithella Hookeriana, x 2/3.— The arrangement
of the flgures is —
1
4
7
10
2
5
8
11
3
6
9
12
1-6, var. Willisiana (C. A. Barber, Xo. 2,515) ; 7-12, var. Bombayensis
( Atgaon). Illustrating the extraordinary polymorphism of the thallus.
Some are only in bud, some in ripe fruit. 1 and 4 are similar forms from
the top of the rock, showing radial symmetry ; the latter is seen from
below, and has only a very short basal stalk. 2 and 3 are from the sides
of the rock, and are drawn out on the lower side. 5 and 6 are prostrate
discoid forms closely attached at all points to the rock. The Atgaon
forms are very complex in shape ; the foot can be seen in 7, 8, 9, 10.
No. 9 is seen from below. No. 11 is in fully ripe fruiting condition, and
the thallus is much whiter than in the rest.
Plate XXVII. is omitted, see above.
Plate XXVIII., Willisia selaginoides, Hydrobryurn lichenoides
Fentonii. — A large stone, over which the water was still flowing to a
depth of about an inch, from the Sholai Aar, Anamalais, January, 1901,
X 7/18. This shows particularly well the habit of these two forms. The
Willisia is in ripe fruit, though the lower parts of the secondary shoots are
still submerged ; in many cases the fruits appear to be stalked, owing to
the falling away of the cortical tissues of the stems. The prostrate branch-
ing habit of the Hydrobryum is easily seen, and the numerous unopened
floral buds at the angles of the branching {cf. PI. XXXI.).
Plate XXIX., Willisia selaginoides.—From the same place, speci-
mens in ripe fruit, X 2/3. These show the 4 -ranked scale leaves and the
sessile fruits ; in some more or less of the cortical tissues have fallen away,
leaving the fruits on stalks.
Plate XXX. — The same. 1 A portion of the crustaceous thallus,
gathered in January from below the water, with secondary shoots (s.s."),
showing the mode in which new ones are formed at this season, X SJ. 2
A small plant from herbarium material collected in October, showing two
axes of the type I imagine to be primary, one broken off at the middle.
The leaves of these are long and narrow, like the tips of the leaves of the
secondary shoots (nat. size, drawn by A. de Alwis). 3 Spathe at top
of shoot, X 4. 4 Open flower at top of shoot, X 4. 5 Ripe fruit, X 4.
6 Open fruit, x 4, showing the way in which the valve bends downwards
after dehiscence, and the double stalk on which the fruit appears to stand
owing to the fall of the cortical tissues. 7 Transverse section of ripening
pericarp, showing inner epidermis with its cuticle (i.e., out.), the scleren-
chyma layers (■'?c/.), and the outer deciduous parenchyma ipar.'), X 600.
OP CEYLON AND INDIA.
m
Plate XXXI., Hydrobryum lichenoides, x 11 18.— Order of figures
1 8 I
2 5
1, 2, and 5 are var. Khaudalense^ gathered on perfectly dry rocks, the
plants quite dead, the fruits mostly open. 8 is var. Bkorense, with its
long-stalked fruits. 4 is var. Khandnlcnse, gathered below the water, to
show {cf. PI. XXVIII.) the habit and the sessile prostrate floral shoots.
In 1 the tips of the leaves are persistent in many of the shoots ; in 2 the
very short-stalked fruits can be made out.
Plate XXXII.' — The same (figs. 1-6). — A seedling of var. keleiisis,
from Hatton, x I, showing cotyledons (cot.), primary axis, origin of the
thallus (th.), and secondary shoots (^.ä-.). 2 Tip of thallus of a young
plant of the same, X 4, showing thallus and secondary shoots, and course
of vascular strands (v.h.). 3 Portion of thallus developing a floral shoot
from a former vegetative shoot, X SJ, showing enlargement of bases of
leaves. 4 Portion of thallus of same, in ripe fruit, X ; the spathe is
split irregularly, and the fruit-stalk stands erect. The shoot at # is
omitted. 5 Stigmas of three flowers of Shillong ianum, x 6, to show
variety of form ; the stamens in section on lower side. 6 Stigmas of
var. Khanclalense, X 6.
H. sessile (figs. 7-9). 7 Portion of thallus in ripe fruit, x 6, to show
branching ; the lobes of the thallus almost touch each other. The shoots
are like those of the other species. 8 Open flower, X showing bracts
long stamens, &c., from above. 9 The same from the side, showing the
perianth segment or staminode, std., and curvature of ovary, &c.
H. olivaceum zeylanicum (figs. 10-18). — 10 Embryo from ripe
seed, X 6, 11 A young seedling, X 6, showing cotyledons, hypocotyl,
hapteron (A.), and commencement of thallus (th.'). 12 A rather older
seedling, X 6, with the first leaves forming, and with more than one
hapteron. 13 The same, from above, to show arrangement of leaves at
growing point, 14 An older seedling, clasping an old capsule, x 8. 15
Tip of a primary axis, X (H, showing arrangement of leaves. 16 Tip.
from above, X 6. 17 Tip of older axis, X 6, with three growing points.
18 Seedling with thallus developing higher up than usual upon the
hypocotyl, X 8.
Plate XXXIII., Hydrobryum olivaceum, dead plants in ripe fruit, x
3/4. — The upper left-hand stone is covered with the var. anamalaiense,
turned to show the ribbed fruits on their short stalks. The others are all
var. griseum, and show the habit well, especially the small plant in the
lower left-hand corner. The large right-hand rock shows the way in
which the thallus in this form often becomes crumpled and ridged. The
secondary shoots can be easily made out, with the large terminal spathe
pointing towards the margin of the thallus and open on the upper side.
In the middle of the large rock, on the left of the chief thallus, two
portions of another plant are colliding with it.
Plate XXXIY., Hydrobryum olivaceum zeylanicum, — 1 A seedling
clasping an old capsule, X 5^. The primary axis has been lost, and the
secondaries are much larger than usual in so small a plant ; th., thallus ;
.v.-v.. secondary shoots ; ^., the old cai>sule stalk. 2 Vertical longitudinal
4(i4 WILL 18 : MOHPHOLOLY OF THE PODü8'I'BMACEÆ
section of growin” margin of thallns, X 320, showing meristem and root-
cap. 0'.6'.). 3 A young plant in May, nat. size ; the primary axis ;
S.S., the secondary shoots ; th., the thallus. 4 Slightly diagrammatic
reproduction of the original drawing of Podostemon G-ardneri (the
primary axis. with its brush of leaves), with the addition of part of
the thallus, and with two secondary shoots ; gathered in August ; nat.
size. 5 Transverse section of thallus, x 320, showing epidermis (<9^?.),
parenchymatous cortex Cp.c.), and a vascular strand (f.b.'). 6 Portion of
thallus, X li, showing stages in the transformation of vegetative into
floral shoots by the enlargement of the bases o f the now prostrate leaves
and the fall of their tips.
Plate XXXV., Hydrohryum olivaceum zeylanicum, a plant in
vegetative condition, gathered in August, X 4/5.— This shows the habit
particularly well, but it should be noted that most of the leaves, both of
the primary shoot and the secondaries, have fallen off, owing to the
specimen having been kept in standing water for some time. If photo-
graphed in the perfectly fresh condition, the thallus would be completely
hidden by the leaves, except at the margins. The primary axis is visible
at the lower side of the picture, and the mode of branching of the thallu
is clearly seen. The photograph shows, what is not visible to the eye, the
actual overlapping of the lobes in one or two places. The younger leaves
of the secondary shoots near the margins can be seen emerging from the
thallus,
Plate XXXVI., Hydrohryum olivaceum zeylanicum (figs, 1-5).— 1
Secondary shoot, X 4, the flower just about to emerge from the spathe
(.çj^.), which has split along the upper side. 2 Open flower, from the side,
X 4. 3 Piece of rejuvenescent thallus, developing from an old portion
which has been re-submerged, X 7. At an the first sign of appearance of
the secondary shoot ; .s*. .9., an older shoot. 4 Rejuvenescence from the side,
X 7. 5 A monstrous thallus, described at p, 391 ; the primary axis
bent down
H. Griffithii (figs. 6-8). — 6 Portion of thallus, near a sinus, with
secondary shoot in floral condition at the angle, X 4. 7 Ovary and stigmas
from side, X 4|. 8 Stigmas of various flowers from same locality, X 8, to
show variety of form {of. PL XXXII., figs. 5, 6).
Farmeria metzgerioides (figs, 9-16). — 9 Embryo from seed, x 8.
10 Young seedling, X 5, showing cotyledons (^cot.') and first development of
thallus {til.') ; /., the foot. 11 Larger seedling, with two leaves, X 5L 12
Older seedling, with first secondary shoot forming (9.9.), X 5J. 13 Seedling
with two thalli, from above, X 5^. 14 Two seedlings starting from the
seeds in a capsule enclosed in the old bracts {hr.), X 5J. 15 Seedling with
two thalli, X 5L 16 The same seedling as in fig. 13, from below.
Plate XXXVII., Farmeria metzgerioides (figs. 1-9).— 1 Young plant, x
5 ; the cotyledons and primary axis are still present, while the thallus has
grown out and branched {hr.), the secondary shoot (9.9.) being on the
posterior side of the branch. 2 A thallus forming flowering shoots in
December, x 4. 3 Thallus with branch {hr.) emerging on anterior side of
secondary shoot : the arrow points towards the apex of the thallus.
4 Stamens, X 5. 5 Conteuts of ovafy. from above ; the thick distal
OF CFYLON AND INDIA.
46f)
placenta (^^.) bears two ovules on the basal side. X 20. 6 The same, from
the side, showing the sex^tum ( sjJt,'). 7 The same, from the upper side, with
the ovules bent down ; s., the stalk bearing the thick placenta. 8 The
fruit, X 20, from the side. 9 Interior of fruit, from the lower side, X 20,
showing the two septa, the stalk of the placenta, and the seeds.
Farmeria indica. — ^10 Plant, from shrivelled rock-dry material, show-
ing thallus (th.) with hapterous discoid feet (/.), X 6.
Plate XXXVIII., Farmeria indica.— The three small stones at the left
show this species, with its branching habit, and small discoid feet, X 3/5.
Farmeria metzgerioides. — The two central stones show this species,
in ripe fruit, the fruits being held firmly against the rock by the persistent
bracts. Tristicha ramosissima. — The large right-hand stone is covered
with the creeping thalli of this plant, on which the discoid hapterous feet
can be clearly seen ( C. A. Barber, No. 2,518). The secondary shoots have
been removed. C/. PI. V.
T5
CEÏIiON
PRINTED AT THE GOVERNMEN'l PRESS
COLOMBO ^
Ann. Perad. I PL IF
PODOSTEMACEÆ AT HAKINDA.
Ann. Perad. I
PL V
TRISTICHA RAMOSISSIMA, Willis.
Ann. Perad.
i
TRISTICHA RAMOSISSIMA, Willis.
Aun.Peracl.l .
Fl.YIl.
Irtli..Ansf.T.I.A7uriie,Ie]piig‘.
J.C.Wdel.
— — v.h.
— t.coL.
— o.coZ.
— p.c.
v.h.
—ep.
TRISTICHA RAMOSISSIMA
iViui .Ferad.l
Pi.V
J.C.Wiel.
TRISTICHA RAAfOSISSIMA.
lit?L. Ans t.irï. A.Fimte, Leipzig .
Ann. P er ad. I.
PI. IX.
1.
J. C.W. del.
TRISTICHA RAMOSISSIMA (1-3)
Iith.ÄrLsty.E.Ä.TuiilceleiDziff.
LAWIA ZEYLANICA (A-IY)
Ann. Perad.
LAWIA ZEYLANICA, Tul.
V: r:!v
Aim.Perad.I.
PI. XI.
■j.C.W. del
lAWIA ZEYLANICA
litTxAiLst.YE.il'Furlce; Leipzig-,
Ann. Perad. I
PL XII
LAWIA ZEYLANICA, Tul.
Aiui.Perafl.T.
pi.xni.
J.C.Wdel.
liidi.Anst.YÎ-A.Turîke
LAWIA ZKTLANICA.
Ann. Perad. I Pl, XII'
PODOSTEMON SUBULATUS, Gardn. ; DICRÆA SPP.
Ann . Pei*acl, I .
JG.-Wdel.
PODOSTEMON SIJBULATUS.
XiftLA-nstv.E.Alimke .Leipzig-.
Aim.Perad.I.
PI. AAH.
•J.CW. del.. lit>L.Aiist;-'j;E.A,ï’unke, Leipzig-.
P O D O S TE MO N B ARB E RI .
Jfc if-.
AmiTerad.I.
PL xm.
DICRÆA EL ON G ATA.
lith inst.v.E. A, Punte, Leipzig.
J.C.W.del.
Ann. Perad. I
PL XIX
DICRÆA ELONGATA, Tul
Ann .Per ad. I.
PI. XX.
D1CRÆA SPP.
J.CW.del.
iif L Anst ,yJ;. A.rim]ce,Leip zij.
12.
TJ.l).
Ann. Peu ad. I PI XXI
DICRÆA WALLICHII, Tul.
r
h
l-'C-
V'.f ‘
• -.v??”;/:.
Aim. Pera d. I
PI. XXII.
DICRÆA STYLO SA.
JC.Wiel.
s.s.
7.
v:b.
"Kj^CitZc. Jt
97
î Ami.PeradJ.
Pl.xxiv:
j.crvv: del.
Lifa.Aiist.x’:E.AJanlîe,Lei2?ig'.
DIGRÆA SPP. (1-6) GRIFFITHELIA (7-10).
i.
Ann. Per ad. I PI. XXF
GRIFFITHELLA HOOKERIANA, Wmg.
^NN. ttraü;
I;
GRIFFITHELLA HOOKERIANA, Wmg.
^
./ ,,-a
A
â
f
PI. XXVIII
WILLISIA SELAGINOIDES, Wmg. ; HYDROBRYUM LICHENOIDES, Kurz.
Ann. Per ad. I PL XXIX
»
WILLISIA SELAGINOIDES, Wmg.
WILLISIA SELAGINOIDES .
Ann. Per ad. I PL XXXI
HYDROBRYUM LICHENOIDHS, Kurz.
y\iui.I’ei-acl.l.
Pi. XXXII
J.C7W^ del'.
HYDROBRYUM SPP.
Lith.Anst.vE.A.Furike, Leipzig.
"■6Q'
Ann. Perad. I PL XXXIII
HYDROBRYUM OLIVACEUM, Tul.
.Viin.PpraJ.I.
Pl.XXXlV.
mTIROBRXmi OLIVACEUXI.
IitiiAast.'v:E,A.T'UTi'ke,,Ieip2ij.
Ann. Perad. I pi_ XXXF
HYDROBRYUM OLIVACEUM, Tul.
Aim.Perad.I.
pi.xx.wr„
J. C .‘W. del . lifh. Mst.v. E . Al'-unlveileipzi^-.
HYDROBRYUXP SPP.(l-8); FARXŒRIA XI. (9-16 ).
Aim.Pcrüd. f.
Pl.XXXMI.
ÜJ.C.W dei.
lltTi. Anst y,E,A.Funi:e, Leipzig
FARMERIA MPFFZGERIOIDES ( 1-9); F. INDICi\( 10) .
PL XXXVllI
FARMERIA SPP. ; TRISTICHA RAMOSISSIMA, Willis.
Vol. I., Pt. I. JUNE, 1901. Price 50 cents.
8d.
Annals
OF THE
Royal Botanic Gardens,
Peradeniya
EDITED BY
J. 0. WILLIS, M.A., F.L.S.,
DIRECTOR.
CONTENTS.
PACE
WILLIS, J. C. — The Royal Botanic Gardens of Ceylon and their History ... 1
WIT.LIS, J. C. — The Royal Botanic Gardens of Ceylon as a Centre for Botanical
Study and Research ... ... ... ... 17
Columbo :
GEORGE J. A. SKEEN, GOVERNMENT PRINTER, CEYLON.
Cotiitott :
UULAU & Co., 37, SOHO SQUARE, LONDON, W.
\^All rights of Reproduction and Translation reserved.']
ANNALS OF THE ROYAL BOTANIC GARDENS,
PERADENIYA.
NOTICE TO SUBSCRIBERS.
The subscription rate is, for regular residents in Ceylon,
Rs. 2*50 per annum, post free, payable in advance to the
Director, Royal Botanic Gardens, Peradeniya ; for residents
in other countries, Rs. 6 per annum, post free, payable in
advance to the above, or eight shillings, payable to Messrs.
Dulau & Co., 37, Soho Square, London, W.
The Annals will appear at irregular intervals, as matter is
ready for publication, probably averaging 250 pages per
annum, and when a volume of convenient size has been
formed the Title Page and Contents will be issued. Indi-
vidual numbers or papers may be purchased from the
Director, Royal Botanic Gardens, Peradeniya, or from
Messrs. Dulau & Co., at prices marked upon them (con-
siderably exceeding the subscription rate).
The “ Circular ” of the Royal Botanic Gardens, containing
more popularly written articles on planting, agricultural,
and botanical topics, may be received with the Annals by
adding cents 50 to the Ceylon residents’, and Re. 1*50, or 2s,
to the foreign residents’ subscription.
NOTICE TO CONTRIBUTORS.
All contributions should be addressed to the Director,
Royal Botanic Gardens, Peradeniya, Ceylon. They must be
clearly written on one side of the paper only (papers in
French or German must be typewritten), and the figures
must be ready for reproduction, and planned so as to pro-
perly fill a plate.
Each contributor is entitled to receive gratis twenty-five
separate copies of his paper, and may, by giving notice
when returning proofs, purchase others at the rate, approxi-
mately, of two cents per page (including covers, if required),
and 20 cents per full-page plate, for each copy, provided not
less than 10 be taken.
Vol. I., Pt. II.
DECEMBER, 1901,
Price Rs. 3.
4s. Od.
Annals
OF THE
OYAL Botanic Gardens,
Peradeniya.
EDITED BY
J. C. WILLIS, M.A., F.L.S.
DIRECTOR.
CONTENTS.
PAGE
HOLTERMANN, C. — Fungus Cultures in the Tropics ... ... 27
WILLIS, J. C. — Note on the Flora of Minikoi ... ... 39
WILLIS, J. C., and GARDINER, J. S.— The Botany of the Maidive Islands... 45
WRIGHT, H. — Observations on Dracæna reflexa ... ... 165
REVIEWS. NOTES. SUPPLEMENT.
®oItnn&tr :
H. C. COTTLE, ACTING GOVERNMENT PRINTER, CEYLON.
Cottiion :
DULAU & CO., 37, SOHO SQUAEE, LONDON,
\^All rights of Bej^ro duct ion and Translation reserred.'\
ANNALS OF THE ROYAL BOTANIC GARDENS,
PERADENIYA.
NOTICE TO SUBSCRIBERS.
The subscription rate is, for regular residents in Ceylon,
Rs. 2-50 per annum, post free, payable in advance to the
Director, Royal Botanic Gardens, Peradeniya * for residents
in other countries, Rs. 6 per annum, post free, payable in
advance to the above, or eight shillings, payable to Messrs«
Dtjlau & Co., 37, Soho Square, London, W.
The Annals will appear at irregular intervals, as matter is
ready for publication, probably averaging 250 pages per
annum, and when a volume of convenient size has been
formed the Title Page and Contents will be issued. Indi“
vidual Numbers or Papers may be purchased from the
Director, Royal Botanic Gardens, Peradeniya, or from
Messrs. Dulau & Co., at prices marked upon them (con-
siderably exceeding the subscription rate).
The ‘‘ Circular of the Royal Botanic Gardens, containing
more popularly written articles on planting, agricultural,
and botanical topics, may be received with the Annals by
adding cents 50 to the Ceylon residents’, and Re. 1*50, or 2s.,
to the foreign residents’ subscription.
NOTICE TO CONTRIBUTORS.
All contributions should be addressed to the Director,
Royal Botanic Gardens, Peradeniya, Ceylon. They must be
clearly written on one side of the paper only (Papers in
French or German must be type written), and the figures
must be ready for reproduction, and planned so as to
properly fill a plate.
Each Contributor is entitled to receive gratis twenty-five
separate copies of his Paper, and may, by giving notice
when returning proofs, purchase others at the rate, approxi-
mately, of two cents per page (including covers, if required),
and 20 cents per full-page plate, for each copy, provided not
less than 10 be taken.
May, 1902.
Vol. I., Pt. III.
'S'&ù.-iStf
Price Rs. 2‘50,
3s. 4d.
Annals
OF THE
Royal Botanic Gardens,
Peradeniya.
0
EDITED BY
J. C. WILLIS^ M.A., F.L.S.
DIRECTOR.
CONTENTS.
PAGE
WILLIS, J. C.' — A Revision of the Podostemaceæ of India and Ceylon ... 181
REVIEWS of Recent Literature ... ... ... 251
NOTES (Analyses of Ceylon Indiarubber. Jak Milk. Pith for Micro-
scopists. Notes of Indian Travel by a Ceylon Botanist. Opening of
Experiment Station. The New Branch Carden at Nuwara Eliya.
Peradeniya Resthouse. Publications. Personal.) ... ... 253
SUPPLEMENT (A Handbook of the Economic Products of Ceylon, by J. C.
Willis and Herbert Wright).
H. C. COTTLE, ACTING GOVERNMENT PRINTER, CEYLON.
Kmtfm :
DULAU & CO., 37, SOHO SQUARE,
[All rights of Reproduction and
ANNALS OF THE ROYAL BOTANIC GARDENS,
PERADENIYA.
NOTICE TO SUBSCRIBERS.
The subscription rate is, for regular residents in Ceylon,
Rs. 2’50 per annum, post free, payable in advance to the
Director, Royal Botanic Gardens, Peradeniya ; for residents
in other countries, Rs. 6 per annum, post free, payable in
advance to the above, or eight shillings, payable to Messrs.
Dulau & Co., 37, Soho Square, London, W.
The Annals will appear at irregular intervals, as matter is
ready for publication, probably averaging 250 pages per
annum, and when a volume of convenient size has been
formed the Title Page and Contents will be issued. Indi-
vidual Numbers or Papers may be purchased from the
Director, Royal Botanic Gardens, Peradeniya, or from
Messrs. Dulau & Co., at prices marked upon them (con-
siderably exceeding the subscription rate).
The “Circular” of the Royal Botanic Gardens, containing
more popularly written articles on planting, agricultural,
and botanical topics, may be received with the Annals by
adding cents 50 to the Ceylon residents’, and Re. 1*50, or 2s.,
to the foreign residents’ subscription.
NOTICE TO CONTRIBUTORS.
All contributions should be addressed to the Director,
Royal Botanic Gardens, Peradeniya, Ceylon. They must be
clearly written on one side of the paper only (Papers in
French or German must be type written), and the figures
must be ready for reproduction, and planned so as to
properly fill a plate.
Each Contributor is entitled to receive gratis twenty-five
separate copies of his Paper, and may, by giving notice
when returning proofs, purchase others at the rate, approxi-
mately, of two cents per page (including covers, if required),
and 20 cents per full-page plate, for each copy, provided not
less than 10 be taken.
Vol. I., Pt. IV. SEPTEMBER, 1902. Price Rs. 10-50.
14s. Od.
Annals
OF THE
Royal Botanic Gardens,
Peradeniya.
EDITED BY
J. Co WILLIS, M.A., F.L.S.
DIRECTOR.
CONTENTS.
FADE
WILLIS, J. 0.' — Studies in the Morphology and Ecology of the Podostemaceæ *
of Ceylon and India ... ... ... ... 267
€oloxnb0 :
H. C. COTTLE, ACTING GOYERNMENT PRINTER, CEYLON.
Conium ;
DTJLAU & CO., 37, SOHO SQUARE, W.
\_All rights of Reproduction and Translation reserved.~\
ANNALS OF THE ROYAL BOTANIC GARDENS,
PERADENIYA.
NOTICE TO SUBSCRIBERS.
The subscription rate is, for regular residents in Ceylon,
Rs. 2’50 per annum, post free, payable in advance to the
Director, Royal Botanic Gardens, Peradeniya ; for residents
in other countries, Rs. 6 per annum, post free, payable in
advance to the above, or eight shillings, payable to Messrs.
Dulau & Co., 37, Soho Square, London, W.
The ‘‘ Annals ” will appear at irregular intervals, as matter
is ready for publication, probably averaging 250 pages per
annum, and when a volume of convenient size has been
formed the Title Page and Contents will be issued. Indi-
vidual Numbers or Papers may be purchased from the
Director, Royal Botanic Gardens, Peradeniya, or from
Messrs. DuLAU & Co., at prices marked upon them (con-
siderably exceeding the subscription rate).
The “ Circular ” of the Royal Botanic Gardens, containing
more popularly written articles on planting, agricultural,
and botanical topics, may be received with the Annals by
adding cents 50 to the Ceylon residents’, and Re. 1*50, or 2s.,
to the foreign residents’ subscription.
NOTICE TO CONTRIBUTORS.
All contributions should be addressed to the Director,
Royal Botanic Gardens, Peradeniya, Ceylon. They must be
clearly written on one side of the paper only (Papers in
French or German must be type written), and the figures
must be ready for reproduction, and planned so as to
properly fill a plate.
Each Contributor is entitled to receive gratis twenty-five
separate copies of his Paper, and may, by giving notice
when returning proofs, purchase others at the rate, approxi-
mately, of two cents per page (including covers, if required),
and 20 cents per full-page plate, for each copy, provided not
less than 10 be taken.
Vol. I., Pt. V.
OCTOBER, 1903.
Price Re. PSO.
2s. Od.
Annals
OF THE
KOYAL Botanic Gardens,
' Peradeniya.
TITLE PAGE and INDEX to Volume I.
SUPPLEMENT (A Handbook of the Economic Products of Ceylon, by J. C.
I Willis and Herbert Wright), co/tibmerZ.
\ ' a ■ .
GEORGE J. A. SKEEN, GOVERNMENT PRINTER, CEYLON.
EDITED BY
J. C. WILLIS, M.A., F.L.S.
DIRECTOR.
CONTENTS.
Conïnjn :
DULAÜ & CO., 37, SOHO SQUARE, W.
\_All rights of Rejiroduction and Translation reserved.']
ANNALS OF THE ROYAL BOTANIC GARDENS,
PERADENIYA.
NOTICE TO SUBSCRIBERS.
The subscription rate is, for regular residents in Ceylon,
Rs. 2*50 per annum, post free, payable in advance to the
Director, Royal Botanic Gardens, Peradeniya ; for residents
in other countries, Rs. 6 per annum, post free, payable in
advance to the above, or eight shillings, payable to Messrs.
Dulau & Co., 37, Soho Square, London, W.
The “ Annals ” will appear at irregular intervals, as matter
is ready for publication, probably averaging 250 pages per
annum, and when a volume of convenient size has been
formed the Title Page and Contents will be issued. Indi-
vidual Numbers or Papers may be purchased from the
Director, Royal Botanic Gardens, Peradeniya, or from
Messrs. Dulau & Co., at prices marked upon them (con-
siderably exceeding the subscription rate).
The “ Circular ” of the Royal Botanic Gardens, containing
more popularly written articles on planting, agricultural,
and botanical topics, may be received with the Annals by
adding cents 50 to the Ceylon residents’, and Re. 1*50, or 2s.,
to the foreign residents’ subscription.
NOTICE TO CONTRIBUTORS.
All contributions should be addressed to the Director,
Royal Botanic Gardens, Peradeniya, Ceylon. They must be
clearly written on one side of the paper only (Papers in
French or German must be type written), and the figures
must be ready for reproduction, and planned so as to
properly fill a plate.
Each Contributor is entitled to receive gratis twenty-five
separate copies of his Paper, and may, by giving notice
when returning proofs, purchase others at the rate, approxi-
mately, of two cents per page (including covers, if required),
and 20 cents per full-page plate, for each copy, provided not
less than 10 be taken.
NOTICE.
Volume I. is now complete, and is published at Rs. 15 or
£1, unbound. A few copies of Parts I. (Ceylon Botanic
Gardens, &c.), III., and IV. (Podostemaceæ, &c.) can still be
had. The following reprints are also for sale at the prices
marked : —
Willis, A Revision of the Podoste-
maceæ of India and Ceylon, 70 pages Re. 1*50 or 2s. Od,
Willis, Studies in the Morphology
and Ecology of the Podostemaceæ
of Ceylon and India, *200 pages with
33 plates * ... ... Rs. 7*50 or 10s. Oc?.
CEYLON PUBLICATIONS
FOR. SAI.E AT THE GOVERNMENT RECORD OFFICE, COLOMBO.
Oriental Literature. , Es. c.
The Mahavvansa : —
Original Pali Text, Part I. ... 7 50
Do. Part II. ... 7 50
Sinhalese Translation, Part I. ... 5 0
Do. Part II. ... 5 0
Wijesinha’s English Translation of Part II.,
with Tumour’s Translation of Part I. ... 7 50
The Mahawansa Tika, with original Pali,
bound in stiff covers ... ... 7 50
Do. do. unbound ... 6 50
Saddharmalankaraya (Sinhalese) ... 2 0
Extracts from the Pu jawaliya (English) ... I 0
Do. do. (Sinhalese)... 0 75
Nitinighanduwa, English ... .j. 1 0
Do. Sinhalese... ... 1 0
Moggallana Panchika Pradipa (Pali) ... 1 0
Abhidhanappadipika, a Dictionary of the
Pali Language ... ... 8 0
The Tesawalamai ... ... 0 50
Lapidarium Zeylanicum ... ... 31 50
Dravidian Comparative G-rammar ... 1 3 0
Glossary of Native Words occurring in
Official Documents (third edition) ... 0 30
Catalogue of Pali, Sinhalese, and Sanskrit
»V Manuscripts in Temple Libraries ... 0 50
viA. Descriptive Catalogue of Sanskrit,
Pali, and Sinhalese Works ...5 0
Mugdhabodha Wyakarana (Sanskrit) ... 5 0
Mukhamatta Dipani (Pali) ... 5 0
Kawsilumina (Sinhalese) ... ... 1 50
Pybus’s Mission to Kandy ... ... 0 50
Papers on the Custom of Polyandry as
practised in Ceylon ... ... 0 15
The Rajavaliya ; or, a Historical Narrative
of Sinhalese Kings (in English and in Sin-
halese) ... ... each 0 75
Archæology.
Dr. Müller’ s Report on Inscriptions of Ceylon : —
Text ... ... ... 5 0
Plates ... ... ... 5 0
Rs. c.
Architectural Remains of Anuradhapura
(with plates), by J. Gl. Smither,
F.R.I.B.A.
In boards ... ... 40 0
In cloth ... ... 60 0
Return of Architectural and Archæological
Remains and other Antiquities in Ceylon... 1 20
Reports on the Archæological Survey of
Ceylon : —
Kegalla District
... 6
0
Anuradhapura
AI.) ...
... 0
50
Do.
(II.) ...
... 1
0
Do.
(III.) ...
... 1
65
Do.
(IV.) ...
... 1
0
Do.
(V.) ...
... 2
20
Do.
(VI.) ...
... 2
0
Do.
(VII.) ...
... 4
0
Natural History.
Report on Brown Scale (or Bug) on Coffee... 1 0
The Green- Scale Bug in connection with the
Cultivation of Coffee. — Observations by
Mr. E. Ernest G-reen (illustrated) ... 1 0
The Flora of Ceylon, by Dr. Trimen : —
Parts III., IV., and V. (with plates) each 20 0
Lepidoptera of Ceylon, in 13 Parts, with
coloured plates ... each Part 14 50
The Kitul Palm and its Uses, by T, B.
Pohath-Kehelpannala ... ... 0 15
Report on the Ceylon Pearl Fisheries ... 1 35
District Manuals.
Mannar, by W. J. S. Boake ...
... 1 0
Uva, by H. White
... 2 .50i
Nuwara Eliya, by C. J. R. Le Mesurier
... 5 0
Vanni Districts, by J. P. Lewis
... 5 0
North-Central Province, by R. W. levers
... 2.50
Rs. c.
Ceylon Civil List ... ... ... ... 1 0
Ceylon Blue Book ... ... ... ... 10 0
Administration Reports (annual volumes) ... ... 10 0
Sessional Papers (annual volumes) ... ... 10 0
♦
TO BE OBTAINED OE H. W. CAVE &j0o., COLOMBO.
Rs. c.
The Ruined Cities of Ceylon. Demy 8vo. Illustrated with collotypes 9 0
Golden Tips : A description of Ceylon and its great Tea Industry.
214 illustrations from photographs by the Author. Demy 8vo. ... 10 0
Kandy and Peradeniya : an illustrated description, with 33 full-
page photogravures. Demy 4to., Roxburgh binding ... 18 0
Nuwara Eliya and Adam’s Peak : an illustrated description of the
Highlands of Ceylon, with 34 full-page photogravures. Demy 4to.,
Roxburgh binding ... .... ...180
♦
PUBLISHED AT THE ‘‘CEYLON OBSERVER” OFFICE.
Rs. c.
Skeen’s G-uide to Colombo (with maps and illustrated) Re. 1'50 and 1 75
Do. to Kandy ( do. ) ...10
Burrows’ Guide to Kandy and Nuwara Eliya ... ... 1 50
Do. to the Buried Cities ... ... ...2 0
V