ae ie 


AES 
ns 


ere 
yt 


Bor 


A eee, 


=e 


at 4 hens ' Vettes eee 
aerate! . oe * uh Mile ts iat 


‘ 
ear 
ays 
vb ete 
‘ 


ae Kiss eT 
Ulatal Pat eh Nee 


‘ 
he 


tata tects 
mw ee 
= 


Caan 


ieNeeeerre oe spartans 


Patty. 


ree 


Bulletin of the 
British Museum (Natural History) 


Botany series Vol8& 1981 


British Museum (Natural History) 
London 1981 


Dates of publication of the parts 


No 1 3 : : : F 5 . 26 February 1981 
No 2 : c : ; : 26 March 1981 
No 3 : : ! 28 May 1981 
No 4 : ; ; : ; : 30 July 1981 


ISSN 0068-2292 


Printed in Great Britain by Henry Ling Ltd, at the Dorset Press, Dorchester, Dorset 


No 1 


No 2 


Contents 
Botany Volume 8 


Page 
The Thelypteridaceae of Sar 
W. A. Sledge : : : : : ; : 1 
Studies in the genus Hypericum L. (Guttiferae) 2. Characters of the 
genus 
N. K. B. Robson ; ; ; : : ‘ : aS 
A revision of the lichen family Thelotremataceae in Sri Lanka 
M. E. Hale , 2 ; : : : : : .) gar 


Vascular plant collections from the Tristan da Cunha group of islands 
E. W. Groves. ; : ; 2 : ; , 303 


— 
o 


e . = : 
fof (emt E 


nih * ir; - 


7 


oe ee 


| tooth tele Wee he 


d f —_ 
y 
- 


Bulletin of the ( 
British Museum (Natural Histo 


The Thelypteridaceae of Ceylon 


W. A. Sledge 


Botany series Vol8 Nol 26 February 1981 


The Bulletin of the British Museum (Natural History), instituted in 1949, is issued in four 
scientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology, and 
an Historical series. 


Papers in the Bulletin are primarily the results of research carried out on the unique and 
ever-growing collections of the Museum, both by the scientific staff of the Museum and by 
specialists from elsewhere who make use of the Museum’s resources. Many of the papers are 
works of reference that will remain indispensable for years to come. 


Parts are published at irregular intervals as they become ready, each is complete in itself, 
available separately, and individually priced. Volumes contain about 300 pages and several 
volumes may appear within a calendar year. Subscriptions may be placed for one or more of 
the series on either an Annual or Per Volume basis. Prices vary according to the contents of 
the individual parts. Orders and enquiries sould be sent to: 


Publications Sales, 
British Museum (Natural History), f 
Cromwell Road, 
London SW7 5BD, 
England. 


World List abbreviation: Bull. Br. Mus. nat. Hist. (Bot.) 


© Trustees of the British Museum (Natural History), 1981 


ISSN 0068-2292 Botany series 
Vol 8 No 1 pp 1-54 
British Museum (Natural History) 
Cromwell Road 
London SW7 5BD Issued 26 February 1981 


The Thelypteridaceae of Ceylon \ 


W. A. Sledge _ ~~ 
Department of Plant Sciences, University of Leeds, Leeds LS2 9JT 


Synopsis 

Thirty-four species of thelypteroid ferns from Ceylon (Sri Lanka) are described and discussed, and keys 
are provided. Amauropelta hakgalensis Holttum, Trigonospora angustifrons Sledge, T. glandulosa 
Sledge and 7. obtusiloba Sledge are newly described, as are Christella papilio var. repens Sledge and 
Metathelypteris flaccida var. repens Sledge. Four new combinations are made: Trigonospora 
caudipinna (Ching) Sledge, T. zeylanica (Ching) Sledge, Stegnogramma pozoi var. mollisssima (Kunze) 
Sledge, and var. petiolata (Ching) Sledge. Thelypteris confluens (Thunb.) Morton is newly reported 
from Ceylon. Evidence for the occurrence of Christella meeboldii (Rosenst.) Holttum and C. 
subpubescens (Blume) Holttum and for the status of the former species is presented. 


Introduction 


The ferns described in this paper were not recognised as a separate group of species by 
nineteenth-century botanists. Classifications then adopted attached overriding importance 
to soral characters, and the similarities of these characters in thelypteroid and dryopteroid 
ferns led to species from both groups being included in the same genus. In Beddome’s 
Handbook to the Ferns of British India, Ceylon and the Malay Peninsula (1883), nearly all 
the ferns described here were referred to Lastrea or Nephrodium, according to whether the 
species were free-veined or had anastomosing veins. A separation based on venation was still 
being employed 70 years later with Lastrea—or Thelypteris—and Cyclosorus as the 
currently used generic names, though the artificial nature of the separation was being 
increasingly recognised. 

Christensen (1911) was the first to draw attention to the characters which distinguish 
thelypteroid from dryopteroid ferns, though it was not until almost 30 years later that family 
rank was assigned by Ching (1940) to the Thelypteridaceae. Cytological information has 
since supplied strong evidence in support of the morphological and anatomical grounds for 
the separation. Within the Thelypteridaceae the range of chromosome numbers is 27-36 
(excluding 33), whereas Dryopteris and its allies are based on 41. The characters which 
distinguish thelypteroid ferns are: the possession of rhizome scales bearing acicular hairs on 
their margins and, usually, on their surfaces; stipes always containing two vascular bundles 
at the base uniting upwards to give a single deeply concave strand; acicular hairs invariably 
present on both the upper surface of the main axis and the pinnae; sinuses between the pinna 
lobes normally closed at the base by a translucent membrane. 

Though the family Thelypteridaceae has been generally accepted by pteridologists, no 
general agreement exists as to its subdivision. Ching (1963) described 18 genera from 
mainland Asia; Holttum (1971) recognised 23 genera in the Old World, whilst Pichi- 
Sermolli (1977) listed 32 genera for the world as a whole. However, generic boundaries, as 
defined in these schemes, are too finely drawn—as in the Hymenophyllaceae—to be 
acceptable to some botanists. Moreover Wood’s (1973) observations have shown that 
correlation between gross morphology and spore morphology is often lacking, especially 
amongst those species groups having n = 36 as a base chromosome number. Morton (1963) 
included all species within a single genus Thelypteris. Iwatsuki (1964) and Smith (1971, 
1973, 1974), who have made special studies of Asiatic and New World species respectively, 
take a similarly conservative view, referring most species to The/ypteris, within which most 


Bull. Br. Mus. nat. Hist. (Bot.) 8 (1): 1-54 Issued 26 February 1981 


2 W. A. SLEDGE 


of the genera of Ching and Holttum are treated as sections or subgenera. Such differences of 
treatment are largely a matter of subjective judgements, but since both the morphological 
and cytological diversity within the family seem to me to be too great to warrant the 
inclusion of all species within a single genus, I have followed Holttum’s classification. 

About half of the species described here were recognised by Thwaites, Beddome and their 
contemporaries, though often under different specific epithets. Some names used by them 
covered more than one species, notably so in the case of Nephrodium molle and Lastrea 
calcarata. Both these names, as employed by Beddome and Thwaites, included several 
different species. The former collective species incorporating Christella parasitica and related 
taxa, was shown by Manton & Sledge (1954) to produce hybrids in nature in Ceylon* (Sn 
Lanka). Subsequent experimental work (Panigrahi & Manton (1958), Ghatak & Manton (1971)) 
produced several hybrid combinations, and the variability and consequent difficulty in naming 
some wild gatherings is doubtless connected in some way with hybridisation. Experimental 
studies with the other collective species have not been made, but the polymorphism of the 
Trigonospora calcarata aggregate in Ceylon and south India, where it has long been a source of 
taxonomic confusion, may well also be due, at least in part, to hybridisation between the several 
taxa which the name embraces. In the following account I recognise seven species of Christella 
and seven species, three of them newly described, of Trigonospora, but further observations in 
the field and more cytotaxonomic studies are required, especially in Trigonospora, before a full 
understanding can be reached. Apart from these genera, one new species is described and two 
species are recorded for the first time from Ceylon. 

The thelypteroid ferns of Ceylon are mostly plants of medium to high elevations 
growing terrestrially in shady forest, partially shaded banks, in moist ground near streams, or 
in open swamps. A few are exclusively low altitude ferns. These include Ampelopteris 
prolifera, Cyclosorus interruptus, Trigonospora glandulosa, and T. zeylanica, all of which 
grow in marshes or by the sides of streams. Some species, such as Amphineuron opulentum, 
A. terminans, Christella parasitica, Macrothelypteris torresiana, Pneumatopteris truncata, 
Pronephrium triphyllum, Trigonospora calcarata, and T. obtusiloba, grow both in the low 
country and in the hills, some of them ascending to 1250 m or more. Another group is 
confined to the higher mountains of the interior, growing mostly between 1500-2200 m. 
This group includes Amauropelta hakgalensis, Christella papilio, Metathelypteris flaccida, 
Parathelypteris beddomei, Pseudophegopteris pyrrhorhachis, Stegnogramma pozoi vat. 
petiolata, and Trigonospora caudipinna. Other species are mainly found at altitudes 
between 500-1500 m. 

Some of the commonest ferns in the country belong to this family. The most widely 
distributed and abundant species are Christella dentata, C. parasitica, Macrothelypteris 
torresiana, Sphaerostephanos arbuscula, S. unitus and Trigonospora obtusiloba. Many 
other species, though more restricted in distribution by altitude or ecology, are abundant in 
suitable terrain over considerable areas. The following species are at present known from 
only a few or single localities: Amauropelta hakgalensis, Ampelopteris prolifera, Christella 
meeboldii, C. subpubescens, C. zeylanica, Sphaerostephanos subtruncatus, Thelypteris 
confluens, Trigonospora angustifrons, and T. glandulosa. 

The geographical distribution of some species is wide. Christella hispidula and Cyclosorus 
interruptus are pan-tropical; Ampelopteris prolifera, Amphineuron opulentum, Christella 
dentata, Macrothelypteris torresiana, Sphaerostephanos unitus, and Thelypteris confluens 
spread from tropical or subtropical Africa eastwards across Asia to the Pacific; Christella 
dentata and Macrothelypteris torresiana also occur as adventives in tropical America. Most 
species are confined to south-east Asia ranging eastwards to varying extents, often to China 
and Japan or through Malesia, whilst Amphineuron terminans and Pronephrium triphyllum 
reach northern Queensland in Australia. 

Of the species with a more restricted range, Christella meeboldii, Pronephrium thwaitesii 
and Pseudocyclosorus tylodes are known only from Ceylon and southern India. Sphaero- 


*The name Ceylon is used for Sri Lanka in order to conform with the author’s account of the Dryopteroid ferns, which 
appeared in Volume 5 (1973). 


THELY PTERIDACEAE OF CEYLON 3 


stephanos subtruncatus occurs elsewhere in south-west India and the Seychelle Islands, and 
S. arbuscula in the Mascarene Islands and Madagascar (with a subspecies in east Africa). 
Trigonospora calcarata was described from Java and is reported also from Sumatra, whilst 
Stenogramma pozoi var. petiolata and Christella zeylanica are known to occur outside 
Ceylon only in Java and the Nicobar Islands respectively. 

Six species are endemic to Ceylon: Amauropelta hakgalensis, Pronephrium gardneri 
(probably now extinct), Trigonospora obtusiloba, T. angustifrons, T. glandulosa and T. 
zeylanica. 

In the citations of specimens under each species the C.P. (Coll. Peradeniya) numbered 
sheets distributed by Thwaites have been listed first, since sets of these are to be found in 
many herbaria and the numbers have been widely quoted. Thwaites rarely indicated, save in 
the most general terms, where his specimens had been collected. Much of his material was 
doubtless gathered and brought in by native collectors. The pencilled localities found on 
many of the sheets at Peradeniya indicate localities whence the species had been recorded, 
and not where the specimen on the sheet had been gathered. Since errors of identification 
occur, the localities are not always reliable. 

Gardner also distributed numbered sheets of Ceylon plants, the ferns in his exsiccatum 
being included under the numbers 1053-1271. These are listed next in the specimen 
citations. Gardner’s numbers are quoted in parentheses after the C.P. numbers in Thwaites’ 
Enumeratio Plantarum ‘Zelaniae (1864), whilst his own collection of Ceylon ferns, giving 
name, place of origin and date of collection on each sheet, is at Cambridge. Other gatherings 
by Gardner at Kew and elsewhere bear numbers only. The specimens of other collectors are 
arranged geographically, with unlocalised gatherings placed at the end. 


Acknowledgements 


I thank the directors and curators of the following herbaria for allowing me access to or for the loan of 
specimens in their charge: 

BM: British Museum (Natural History), London; CGE: Botany School, University of Cambridge; E: 
Royal Botanic Gardens, Edinburgh; K: Royal Botanic Gardens, Kew; L: Rijksherbarium, Leiden; P: 
Muséum National d’Histoire Naturelle, Paris; PDA: Botanic Gardens, Peradeniya; US: United States 
National Museum, Smithsonian Institution, Washington, U.S.A.; WRSL: Museum of Natural History, 
Wroclaw, Poland. 

Abbreviations used in the citation of specimens are those adopted in the /ndex Herbariorum 
(Holmgren & Keuken, 1974). 

Professor R. E. Holttum’s unrivalled knowledge of thelypteroid ferns has been freely available to me 
whenever I sought his opinion and | thank him for his valuable comments and unfailing helpfulness. I 
am grateful to Mr J. W. Grimes for supplying scanning electron microscope photographs of spores, to 
Miss Antonia Lant for making the drawings, and to MrG. A. Shaw for typing my manuscript. 


References 


Beddome, R. H. 1883. Handbook to the Ferns of British India, Ceylon and the Malay Peninsula. 
Calcutta. 

Ching, R. C. 1940. On natural classification of the family ““Polypodiaceae.”’ Sunyatsenia 5 : 201-268. 

— 1963. A reclassification of the family Thelypteridaceae from the mainland of Asia. Acta phytotax. 
sin. 8 : 289-335. 

Christensen, C. 1911. Ona natural classification of the species of Dryopteris. In Rosenvinge, L. K. (Ed.) 
Biol. Arb. Tilegn. Eug. Warming : 73-85. Kobenhayn. 

Ghatak, J. & Manton, I. 1971. Further cytological and taxonomic observations on some members of 
the Cyclosorus parasiticus complex. Br. Fern Gaz. 10 : 183-189. 

Holmgren, P. K. & Keuken, W. 1974. Index Herbariorum. Part I. The Herbaria of the World. 6th ed. 
[Regnum veg. 92] Utrecht. 

Holttum R. E. 1971. Studies in the family Thelypteridaceae III. A new system of genera in the Old World. 
Blumea 19 : 17-52. 

Iwatsuki, K. 1964. Taxonomy of the thelypteroid ferns with special reference to the species of Japan 
and adjacent regions. III. Classification. Mem. Coll. Sci. Kyoto Univ. B, 31, : 11-40. 


4 W. A. SLEDGE 


Manton, I. & Sledge, W. A. 1954. Observations on the cytology and taxonomy of the pteridophyte 
flora of Ceylon. Phil. Trans. R. Soc. B, 238 : 127-185. 

Morton, C. V. 1963. The classification of Thelypteris. Am. Fern J. 53 : 149-154. 

Panigrahi, G. & Manton, I. 1958. Cytological and taxonomic observations on some members of the 
Cyclosorus parasiticus complex. J. Linn. Soc. (Bot.) 55 : 729-743. 

Pichi-Sermolli, R. E. G. 1977. Tentamen pteridophytorum genera in taxonomicum ordinem redigendi. 
Webbia 31 : 313-512. 

Smith, A. R. 1971. Systematics of the neotropical species of Thelypteris section Cyclosorus. Univ. Calif. 
Publ. Bot. 59 :; 1-136. 

— 1973. The Mexican species of Thelypteris subgenera Amauropelta and Goniopteris. Am. Fern J. 
63 : 116-127. 

— 1974. A revised classification of Thelypteris subgenus Amauropelta. Am. Fern J. 64: 83-95. 

Wood, C. C. 1973. Spore variation in the Thelypteridaceae. In A. C. Jeremy, J. A. Crabbe & B. A. 
Thomas (Eds), The phylogeny and classification of the ferns. J. Linn. Soc. (Bot.) 67 (Suppl. 1): 191-202. 


Taxonomy 


1 Veins free. bes : ey ene? 29. co pmeteaaey ate) iteless taht Z 
- Veins anastomosing . Re 10 


2(1)  Rhachis reddish-brown, fonds BiDiiate, or dest piauinatitide cei exindusiate 
Pseudophegopteris pyrrhorhachis (p. 5) 


- Rhachis not reddish-brown; sori round and indusiate, or elongate and exindusiate 3 
3(2) | Fronds bipinnate, tripinnatifid . . E : Macrothelypteris torresiana (p. 7) 
- Fronds pinnate or bipinnatifid . ; ; F : : ; 3 : ‘ : ; 4 


4(3) Pinnae pinnatifid with lobed pinnules, veins forked, not reaching the margins 
N Metathelypteris flaccida (p. 8) 


- Pinnae lobed to pinnatifid with entire pinnules, veins simple, reaching the margins 5 

5(4) Sori elongate, exindusiate, pinnae shallowly lobed J Stegnogramma pozoi var. 
petiolata (p. 49) 

- Sori round, indusiate, pinnae pinnatifid . i : : : : : : ; : 6 

6(5) Rhizome creeping v 

- Rhizomeerect . : : oe ae 8 

7(6) Under surface of pinnae ith veatiered sends aid hay costae . Parathelypteris 


beddomei (p. | 1) 

- Under surface of pinnae eglandular, costae bearing scales . Thelypteris confluens (p. 12) 
8(6) | Lower pinnae not reduced orifsonotabruptly . ROAR 3 : . : : 9 

- Lower pinnae abruptly reduced to tubercles on the stipe Pseudocyclosorus tylodes 

(p. 25) 

9(8) | Fronds decrescent, under surface of pees bearing hooked hairs, indusia minute 

Amauropelta hakgalensis (p. 9) 

- eronds not deereseeat ates Sapmce a pinnae without hooked hairs, indusia evident 

Trigonospora (key on p. 15) 


10(1) Sorielongate, exindusiate . , " : , ; , 3 ; al aa: ae 11 
- Sori round, indusiate . P : 3 ; ; ; 3 ; : ; , 14 
11(10) Proliferating buds present on hate bie , , ; Ampelopteris prolifera (p. 14) 
= Proliferating buds absent fromrhachis . . Lee Mie 12 
12(11) Up to 9 pairs of pinnae, no hooked hairs on under atte at pinnge Stegnogramma 
pozoi var. petiolata (p. 49) 
- One to four pairs of pinnae, lower surface of pinnae bearing hooked hairs : : 13 
13(12) Fronds with 1-2 pairs of pinnae, terminal pinna entire . Pronephrium triphyllum 
(p. 47) 
- Fronds with 2-4 pairs of pinnae, terminal pinna pinnate or pinnatifid Pronephrium 


thwaitesii (p. 47) 


THELYPTERIDACEAE OF CEYLON S 


Ma(iO)Atleast3paisofveimsanastomosing . . . «© «| «© 2 as . 15 
- 1-2 pairs of veins anastomosing ; : SAGA iw We 16 
15(14) Margins of pinnae shallowly lobed, 3-5 pairs of veins dielnitekinare Pronephrium 
articulatum (p. 45) 
- Margins of pinnae crenate, 7 pairs of veins anastomosing. Pronephrium gardneri 
(p. 46) 
MEAIGereCt, Cl lm 17 
- Rhizome creeping : ee a, 8) Ee ee Bh 20 
17(16) Pinnae eglandular beheath od Soyagh, ev Vs Dev G o9e Paras Aaa ae 18 
- Ennaewithivellowglandsibeneathi...4 — » = « a wi ws te 19 
hcloeemnaeglabrousbeneath . . .  . .«. 4 Pneumatopteris truncata (p. 42) 
- Pinnaehairybeneath. .. nie iy ary ee Christella hispidula (p. 33) 
19(17) Lower pinnae progressively rediieed. Pe Sphaerostephanos arbuscula (p.43) 
- Lower pinnae abruptly reduced OTs Sphaerostephanos subtruncatus (p. 44) 
20(16) Lower pinnae abruptly reduced Beat Uso Sphaerostephanos unitus (p. 44) 
- Eawenpinnae notorgraduallyreduced . . .°.0 s «© « ».« « « 21 
micimeowenpinnacenotshortened “=> 9% . . « %%  . «6 « « «6% 22 
- Lower pinnae gradually decreasinginsize . . . ‘Christella (key on p. 29) 
22(21) Costae glabrous or sparsely scaly beneath Tee Cyclosorus interruptus (p. 13) 
- Costae hairy and often glandular beneath et id eae ee. MEAL coat 23 
23(22) Lower pinnae narrowed at their bases, not deflexed _.. t : ; . : : 24 
- Lower pinnae not narrowed at their bases, usually deflexed . Christella parasitica 
(p. 29) 
24(23) Margins lobed 3 to costa, at most one pair of veins fusing, sometimes basal veins 
free . fs Ne a. Amphineuron opulentum (p. 27) 
- Margins lobed } to } to costa, 12 pairs of veins fusing . Amphineuron terminans (p. 28) 


I. PSEUDOPHEGOPTERIS Ching 


in Acta phytotax. sin. 8 : 313 (1963). Holttum in Blumea 17 : 12 (1969). 


Caudex erect or creeping; basal scales thin, brown, with scattered superficial hairs but 
lacking conspicuous marginal or apical hairs; stipe and rhachis glossy, usually red-brown in 
colour; fronds elongate, bipinnate with the lower pinnae + reduced and more widely spaced; 
pinnae opposite or nearly so, the pinnules adnate and usually + deeply lobed and often 
connected by a narrow wing; upper surface of costae not grooved; veins usually forked, not 
reaching the margins, tips + thickened; unicellular acicular and/or capitate hairs and much 
reduced filiform scales present on lower surfaces of rhachis and pinna rhachis; sori 
exindusiate, round or slightly elongated, sporangia hairy or glabrous; spores usually with a 
slightly raised surface reticulum, lacking a winged perispore. n = 31. 

Holttum has monographed the genus in Blumea 17: 12-25 (1969). He recognises 20 
species, mostly from south-east Asia, but with outlying species, often very restricted in 
distribution, ranging from St Helena eastwards to Hawaii and Samoa. 


1. Pseudophegopteris pyrrhorhachis (Kunze) Ching 


in Acta phytotax. sin. 8: 315 (1963).—Polypodium pyrrhorhachis Kunze in Linnaea 24: 257 (1851). 
Type: India, Nilgiris, Weigle-Schaeffer 6 (2B, not seen).—Lastrea pyrrhorhachis (Kunze) Copel., 
Gen. Fil. : 139 (1947) p.p.—Macrothelypteris pyrrhorhachis (Kunze) Pic. Ser. in Webbia 24 : 716 (1970). 

Polypodium distans D. Don, Prod. Fl. Nepal. : 2 (1825), non Kaulf. (1824). Baker in Hook. & Baker 
Syn. Fil.: 308 (1867) p.p. Type: Nepal, Wallich (not located)—Phegopteris distans (D. Don.) 
Mettenius in Abh. senckenb. naturforsch. Ges. 3: 16 (1858). Bedd., Handb. Ferns Brit. Ind. : 292 
(1883). 


6 W. A. SLEDGE 


Nephrodium microstegium Hook., Spec. Fil. 4: 119, t. 250 (1862). Type: Khasya, J. D. Hook. & Thomson 
(K).—Lastrea microstegia (Hook.) Bedd., Ferns Brit. Ind. : t.39 (1865). 

Polypodium paludosum sensu Bedd., Ferns South. Ind. :t. 168 (1863), non Blume (1829).—Thelypteris 
paludosa sensu K. Iwats. in Acta phytotax. geobot. Kyoto 19 : 11 (1961). Mem. Coll. Sci. Kyoto Univ. B, 
31 : 139 (1965) p.p. 

Polypodium brunneum Wall., Cat. no. 333 (1829), nom. nud.—Dryopteris brunnea Christensen, Index 
Fil. :255 (1905), nom. nud.—Thelypteris brunnea Ching in Bull. Fan meml Inst. Biol. (Bot.) 6: 269 

(1963), nom. nud. 


Rhizome suberect or short-creeping, stipes 20-60 cm long, straw coloured below becoming 
suffused with red upwards and the rhachis always reddish-brown in colour; lamina 40-80 cm 
long, bipinnate of deeply bipinnatifid with about 20 pairs of free sessile opposite pinnae, the 
lower 2-3 pairs reduced and the lowermost pair more widely spaced on the rhachis; largest 
pinnae about 15 x 4 cm divided nearly or quite to the pinna-rhachis, the pinnules adjacent to 
the main rhachis pinnatifid half way to the costule, becoming progressively less incised then 
entire distally, all with broad bases and contiguous or joined by a narrow wing on the axis of 
the pinna or quite free, apex blunt or subacute, veins pinnate in the lobes, simple or forked, 
pellucid; upper surface short-hairy on costa and costules, lower surface with scattered, 
longer, spreading hairs on the costa and veins and about the margins of the lobes; texture 
herbaceous; sori small, exindusiate, mostly near the extremity of the basal acroscopic forks of 
the veins. 


Collections: Thwaites C.P. 1288 (BM; CGE; K; P; PDA). Nuwara Eliya, moist open places, Oct. 1845, 
Gardner 1151 (CGE; BM; K). Adam’s Peak, 1950 m, 14 Dec. 1950, Sledge 621 (BM). Nuwara Eilya, 
10 May 1906, Matthew 637 (K). Same locality, Freeman 287, 288, 289, 290 (BM). Same locality, 21 
Jan. 1908, Bicknell (P). Same locality, Bradford (CGE). Moon Plains, Nuwara Eliya, 1800 m, 23 Dec. 
1950, Sledge 712 (BM). Adam’s Peak, 14 Feb. 1908, Matthew (K). By track through jungle near Horton 
Plains, 2100 m, 19 Dec. 1950, Sledge 681 (BM). Forest above Le Vallon tea estate, 1350 m, 9 Feb. 
1954, Sledge 1113 (BM). Blackwood Forest, Welimada-Haputala road, Badulla District, by stream, c. 
1500 m, 18 Nov. 1976, Faden 76/345 (K). Unlocalised: Walker (K). Robinson 190 (K). Wall ex herb. 
Hope (P). 

In marshy places and by stream banks at elevations of 1250-2100 m. 


Southern India and Ceylon; northern India from Kashmir to Assam, west China, and 
Vietnam (Tonkin). 


A frequent species at high elevations, growing in the open or in lightly shaded places. 
Readily recognised by its ruddy-coloured rhachis and opposite pinnae with the basal 
pinnules contiguous with or overlapping the frond axis. 

Two cytotypes exist in Ceylon, a tetraploid and a hexaploid. The tetraploid, which 
appears to be the more frequent of the two, differs from the hexaploid in its somewhat 
narrower and more deeply lobed pinnules which, in the middle and distal pinnae, have more 
acute apices. Pinnae of the two types are illustrated in Manton & Sledge, Phil. Trans. R. Soc. 
B, 238 : 162, fig. 13 (1954). 


Il. MACROTHELYPTERIS (H. Ito) Ching 


in Acta phytotax sin. 8 : 308 (1963). 


Caudex short-creeping or erect; scales at base of stipe + thickened at the base, with marginal 
and surface acicular or capitate hairs; fronds bipinnate—tripinnatifid with + adnate 
pinnules, lowest pinnae not or little reduced; upper surfaces of costae not grooved; scales on 
rhachis narrow with thickened base and acicular hair-tip, hairs on lower surface of frond 
long, slender and multicellular; veins usually branched, not reaching margins; sori small, 
indusiate but indusium very small, sporangia usually bearing capitate hairs near annulus, 
spores with a + winged perispore. n= 31. 

Nine species: from Mascarene Islands throughout warmer parts of Asia eastwards from 
Japan to Australia (Queensland); Pacific Islands. 

An account of the species has been given by Holttum in Blumea 17 : 25-32 (1969). 


THELYPTERIDACEAE OF CEYLON 7 
1. Macrothelypteris torresiana (Gaudich.) Ching 


in Acta phytotax. sin. 8 : 310 (1963).—Polystichum torresianum Gaudich. in Freyc., Voy. Bot. : 333 (1828). 
Type: Mariana Island, Gaudichaud (P).—Thelypteris torresiana (Gaudich.) Alston, in Lilloa 30: 111 
(1960). K. Iwats. in Mem. Coll. Sci. Kyoto Univ. B, 31 : 151, 153 (1965). 

Aspidium uliginosum Kunze in Linnaea 20:6 (1847). Type: cult. Hort. Bot. Leipzig, origin Java 
(formerly LZ, now destroyed).—Dryopteris uliginosa (Kunze) Christensen, /ndex Fil. suppl. II: 100 
(1934).—Thelypteris uliginosa (Kunze) Ching in Bull. Fan. meml Inst. Biol. (Bot.) 6 : 342 (1936). 
Holttum, Rev. Fl. Malaya 2 : 241 (1955). 

Polypodium tenericaule Hook. in Hooker’s J. Bot. 9: 353 (1857). Type: China, Alexander (K).— 
Lastrea tenericaulis (Hook.) Moore, Index Fil.:99 (1858). Bedd., Handb. Ferns Brit. Ind. : 266 
(1883).—Nephrodium tenericaule (Hook.) Hook., Spec. Fil. 4: 142, excl. t. 269 (1862) p.p.— 
Aspidium tenericaule (Hook.) Thwaites, Enum. Pl. Zeyl. : 393 (1864). 

Lastrea setigera sensu Bedd., Ferns South. Ind. Correct. p. i (1864) p.p., errore ‘L. flaccida’ in t. 99 
(1863), non Cheilanthes setigera Blume (1828).—Nephrodium setigerum sensu Hook. & Baker, Syn. 
Fil. : 284 (1867) p.p.—Dryopteris setigera sensu Christensen, /ndex Fil. : 292 (1905) p.p. 


Rhizome short-creeping; stipes glaucous when fresh, stramineous when dry, short or long, 
base more or less fleshy and clothed with narrow, brown, hairy scales; distal parts of stipe and 
rhachis glabrous save in the grooved upper surface; /amina bipinnate-pinnatifid or 
subtripinnate, varying widely in size, in small fronds 20 x 12 cm in large fronds four times as 
big, deltoid-ovate in outline, the largest pinnae up to 30 x 15 cm acuminate, lower pinnules 
free the rest and those in upper part of frond adnate to a narrow wing on the pinna axis; 
pinnules acuminate divided to a narrow wing on each side of the costa, the segments oblong, 
blunt-ended, dentate to deeply lobed, glabrous above save for short, antrorse, white hairs on the 
raised costa, lower surface of costae and costules bearing scattered, long, white, mostly 
multicellular hairs and whole undersurface sprinkled with short, capitate, gland-like hairs; 
veins forked, not reaching the margins of the segments, distal parts thickened and 
conspicuous on upper surface; texture herbaceous; sori small, one to each lobe and slightly 
nearer to main vein than margin; indusia very small, membranous, usually bearing a few 
hairs. 


Collections: Thwaites C.P. 1286 (BM; CGE; K; P; PDA.). Thwaites C.P. 1365 (PDA). Rambodde, on 
shady banks, June 1845, Gardner 1150 (BM; CGE; K). Kaduganawa, margins of forests, Oct. 1846, 
Gardner 1222 (CGE; K). Ramboda Pass, 960 m, 17 Dec. 1950, Sledge 656 (BM). Same locality, c. 
925m, 2 Jan. 1977, Faden 77/37 (K). Corbet’s Gap, roadside through secondary jungle, 1290 m, 9 
Dec. 1950, Sledge 572 (BM). Between Hakgala and Ambawela, 1650 m, 25 Dec. 1950, Sledge 788 
(BM). Between Madugoda and Weragamtota, roadside bank through jungle, 750 m, 9 Jan. 1954, Sledge 
949 (BM). Badulla, Freeman 259, 260, 261 (BM). Near Badulla, roadside bank, 29 Dec. 1950, Sledge 
780 (BM). Near Urugala on Kandy-Mahiyangana road, c. 650 m, 24 Dec. 1976, Faden 76/561 (K). 
Opanaki, Kelani Valley near Ratnapura, bushy roadside bank, 5 Jan. 1951, Sledge 801 (BM). 
Manawella, 160 m, Sab. Province, Jan. 1954, Schmid 1067 (BM). Unlocalised: Walker (K). Robinson 
(K). Mrs Chevalier (BM). Alston E71 (PDA). 


Plentiful in the interior in open grassy places or in light shade, from sea level to 1750 m. 

Tropics of the Old World from Madagascar eastwards to southern China, Japan and 
Hawaii and southwards through India and Malaysia to north-eastern Australia and 
Polynesia. Adventive in New World. 


Macrothelypteris torresiana is one of the commonest thelypteroid ferns in Ceylon. It 
varies widely in size, large fronds reaching two metres in height with stipes 120 cm long; the 
largest pinnae of such plants exceed the whole lamina of small plants, wherein only the 
lowest pinnae may be fully pinnate. Normally its subtripinnately dissected fronds bearing 
long, septate, hyaline hairs on the under surface of the pinnae are sufficient to distinguish it 
from all the other thelypteroid ferns. In the field it is easily recognised by its glaucous stipes. 
An insignificant indusium is normally present, but sometimes it is entirely lacking. 


8 W. A. SLEDGE 


III. METATHELYPTERIS (H. Ito) Ching 
in Acta phytotax. sin. 8 : 305 (1963). 


Caudex usually erect; fronds small, pinnate with deeply lobed pinnae or bipinnate with 
adnate pinnules, lowest pinnae not or little reduced; upper surface of costae not grooved; 
veins free, often forked, not reaching the margins; lower surface of pinnae with unicellular, 
acicular and/or short capitate hairs also short multicellular reduced scales; sori indusiate, 
sporangia without hairs, spores dark with thick wings or raised bands. n = 31, 35, 36. 

About 12 species: India and Ceylon to south China and Japan; Malesia; Solomon Islands and 
one species in Sao Tomé and Madagascar. 


1. Metathelypteris flaccida (Blume) Ching 


in Acta phytotax. sin. 8:306 (1963).—Aspidium flaccidum Blume, Enum. Pl. Jav.. 161 (1828). 
Type: Java, Burangrang, Blume.—Lastrea flaccida (Blume) Moore, Index Fil. : 92 (1858). Bedd., 
Ferns South. Ind.:t. 250 (1864); Handb. Ferns Brit. Ind. : 244 (1883).—Nephrodium flaccidum 
(Blume) Hook., Sp. Fil. 4: 133, t.263 (1862); Syn. Fil. : 274 (1867).—Dryopteris flaccida (Blume) 
Kuntze, Rev. Gen. Pl. 2: 812 (1891); Christensen, Index Fil. : 266 (1905).—Thelypteris flaccida 
(Blume) Ching in Bull. Fan mem. Inst. Biol. (Bot.) 6 : 336 (1936). 


Rhizome erect, fronds tufted; stipes up to 30 cm long, slender, black at the base which is 
sparsely scaly, stramineous above, more or less hairy; lamina lanceolate or oblong-lanceolate 
20-40 x 7°5-15 cm, pinnate with deeply pinnatifid pinnae; pinnae subopposite below and 
sometimes throughout, sessile, basal ones shortened and often deflexed, middle ones 
5-10 x 1-5-2 cm, base truncate apex acuminate, pinnatifid to a narrow wing between the 
pinnules; pinnules oblong or linear-oblong, obtuse to subacute at the apex, cut about half 
way to the costa into 4-6 pairs of slightly falcate rounded lobes with ciliate margins; rhachis 
hairy, winged in the distal part; texture softly herbaceous; both sides of pinnae hairy, the 
lower surfaces usually with many long spreading white hairs especially on the costa, costules 
and veins but sometimes only sparsely hairy; veins forked, not reaching the margins; sori 
medial round, at or near the apex of an anterior veinlet, indusium small, membranous, 
margins fringed with hairs. 


Collections: Thwaites C.P. 3802 (BM; CGE; K; P; PDA). Thwaites C.P. 1365 (BM; CGE; K; P) In 
woods at Nuwara Eliya, Sept. 1844, Gardner 1152 (CGE; K). Same locality: Mrs Chevalier (BM); 
Freeman 238, 239, 240, 241 (BM); 1950 m, 11 March 1954, Schmid 1506 (BM). Adams Peak, 14 Feb. 
1905, Matthew (K). Same locality, north slope at 1950 m, 14 Dec. 1950, Sledge 622 (BM); Moon (BM). 
Peacock Hill, Pussilawa, Beckett 71 (BM). Corbets Gap, secondary jungle, 1320 m, 9 Dec. 1950, Sledge 
569 (BM). Jungle between Pattipola and Horton Plains, 1950 m, 20 Dec. 1950, Sledge 672 (BM). 
Horton Plains, 2270 m, 7-8 March 1954, Schmid 1371 (BM). Hakgala, by jungle steam, 1650 m, 27 
Dec. 1950, Sledge 743 (BM). Road between Hakgala and Nuwara Eliya, 1620 m, 27 Dec. 1950, Ballard 
1267 (K). Moon Plains, Nuwara Eliya, wet ground in secondary jungle, 1800 m, 23 Dec. 1950, Sledge 
718 (Bm). By the Ramboda Pass-Maturata track, 1890 m, 17 March 1954, Sledge 1352 (BM). 
Tonacombe Estate, Namunukula, 1350 m, 23 Feb. 1954, Sledge 1180 (BM). Near Ambawela junction 
on Hakgala-Pattipola road, on banks of streams and in roadside ditches, 2 Jan. 1977, Faden 77/20 (K). 
Unlocalised : Walker (K; P; PDA). Wall (K; P). Robinson 158 (K). 


In the higher parts of the interior; frequent by steams and in damp ground in forests. 
Southern India, Ceylon, northern India, southern China, and Java. 
var. repens Sledge, var. nov. 


A typo specie differt: rhizoma late repens, ramosum, frondibus plus minusve approxi- 
matis, stipitibus basibus curvatis. 


Collections: Kandapola, nr. Nuwara Eliya, 1800 m, 19 March 1954, Sledge 1323 (BM, holotype). 
Horton Plains, road to World’s End on roadside banks in forest, +2090 m, 15 Nov. 1976, Faden 
76/284 (K). 


THELYPTERIDACEAE OF CEYLON 9 


Thwaites (in Enum. Pl. Zeyl.:393 (1864)) cites C.P. 1365 as Aspidium tenericaule 
(= Macrothelypteris torresiana); one of the PDA sheets and another at Paris so numbered are 
that species, but 12 other sheets examined are all Metathelypteris flaccida. 

My Kandapola plants with wide-creeping and branching rhizomes are so entirely different in 
habit from the normal erect-growing plant that I assumed that they were specifically distinct. 
Faden’s gathering from Horton Plains is clearly the same. I can find no character however 
which would serve to distinguish fronds detached from plants with creeping and 
erect rhizomes. 

A parallel case is that of Polypodium late-repens, described by Trotter in Hope as a new 
‘ species, largely on the basis of its wide-creeping habit as opposed to the erect-growing P. 
distans D. Don (= Pseudophegopteris pyrrhorhachis). However, Holttum (in Blumea 17 : 24 
(1969)) does not recognise the taxon, even at a varietal level, on the grounds that he ‘can see 
no distinction between specimens lacking rhizomes’. Christella papilio also has both 
creeping and erect forms, the detached fronds of which are likewise inseparable (p. 38). 

Manton found both diploid and tetraploid plants of Metathelypteris flaccida in Ceylon, the 
tetraploid being collected at Horton Plains. I am unable to say however if the tetraploid plant 
had a creeping rhizome, as is the case in Christella papilio. 


IV. AMAUROPELTA Kunze 


in Farnkr. 1: 86, 109, t.51 (1843). 


Caudex erect or a few species with creeping rhizomes; fronds pinnate with deeply pinnatifid 
pinnae, decrescent below, attenuate upwards, aerophores often present at bases of 
pinnae + swollen; veins simple, free, the basal ones passing to the margins above the base of 
the sinus between adjacent segments, sessile glands sometimes present on lower surface, 
short-stipitate, often coloured hairs and acicular or uncinate hairs also often present; sori 
usually supramedial, indusia small, often glandular or hairy, sometimes absent; sporangia 
short-stalked, bearing neither hairs nor glands; spores wingless with a very fine raised 
reticulum. n= 29. 

About 200 species: mostly from tropical and subtropical America from Mexico to Chile 
and Northern Argentina; eight species in Africa, Madagascar and the Mascarene Islands and 
one in Hawaii. Not previously recorded from Asia. 


1. Amauropelta hakgalensis Holttum, sp. nov. (Fig. 1) 


‘A. bergianae (Schlechtendal) Holttum et A. oppositiformi (Christensen) Holttum affinis; a 
priore differt: planta minore, pinnis inferioribus basi non angustatis, pagine inferiore pilis 
capitatis etiam pilis acicularibus hamatis intermixtis vestita; a posteriore differt: pinnis 
fere ad costam lobatis, pagina inferiore pilis hamatis praedita, glandulis subsessilibus rubris 
destituta.’ (Holttum, in /itt.). 


Caudex short, erect; scales thin, to 7 x | mm, bearing a few capitate hairs. Stipe 8-15 cm 
long, with spreading pale hairs almost | mm long. Lamina 40 cm long, pinnae 16 pairs (P. 
220 has lamina 62 cm, pinnae 23 pairs); about 6 pairs of lower pinnae gradually reduced and 
more widely spaced, lowest 3 mm long, base of upper reduced pinnae almost symmetrically 
truncate, not auricled. Largest pinnae 4°5cm long, 1°3-1:5cm wide at base (P. 220, 
5°5 x 1-7 cm), evenly attenuate from truncate base to apex, lobed to within 0°5 mm from 
costa at base, less deeply distally, lobes not falcate, slightly oblique, edges of larger ones 
sinuate; costules 3 mm apart (P. 220 to 4 mm); veins 6-7 pairs (P. 220, 8-9 pairs); lower 
surface of rhachis with spreading pale hairs more than | mm long, of costae with similar 
hairs and shorter ones, distal shorter ones hooked, also short capitate hairs; costules as costae 
with shorter hairs, some hooked; veins and surface between veins bearing short erect hairs, 
hooked and capitate; upper surface of rhachis bearing pale spreading hairs | mm long, with 
shorter ones and capitate hairs in groove, costae with straight hairs 0°7 mm, veins and surface 
with sparse shorter acicular hairs, not hooked. Sori supramedial; indusia very small bearing a 


10 W. A. SLEDGE 


oe 


an X 


sy 4 i We SIONS f 
A ae a YS wG on 
X \ SR “Sls oe 
FpolOl © PO 


Fig. 1 Amauropelta hakgalensis Holttum: A-plant (Holttum S.F.N. 39169, holotype), x 0°5; 
B-part of pinna, x 8; C—under surface of pinna lobe to show sorus, acicular, hooked and capitate 


hairs, x 30. 


few acicular and capitate hairs; spores typical of the genus with a complex fine two- — 


dimensional reticulum. 


Collections: Shrubby slopes above Hakgala Botanic Garden, 1670 m, 23 Dec. 1950, Holttum S.F.N. 
39169 (SING, holotype; K, isotype). Same locality and date, cult. Kew, J. Manton P. 220 (BM). 
Hakgala Bot. Gardens and slopes of Hakgala Mt., 1720-1820 m, 14 Nov. 1976, Faden 76/272 (K). 
Endemic? or possibly introduced, but not referable to any other known species of 
Amauropelta. 


= 


: 


THELYPTERIDACEAE OF CEYLON 11 


This is the fern recorded in Manton & Sledge (Phil. Trans. R. Soc. B, 238 : 137 (1954)) as 
‘P. 220 Thelypteris n.sp.’. Being still unable to match the material with any Indian or 
Malayan fern, specimens were sent to Dr Holttum who recognised it as a species of 
Amauropelta. No species of Amauropelta has hitherto been recorded from Asia, and since 
Holttum reported that the Hakgala fern does not agree with any of the few species known to 
occur in Africa, Madagascar or the Mascarene Islands, he has described the fern, at my 
invitation, as a new species. Meanwhile Dr Faden visited Ceylon and, at our request, 
searched for the new species on the hill slope above the Hakgala Botanic Garden and had 
little difficulty in refinding the plant in two or three widely separated localities. Its 
occurrence in the proximity of a Botanical Garden puts it under suspicion of being an 
introduced species, presumably from America, since it does not agree with any known 
species from the Old World. Accordingly, I sent a gathering to Alan R. Smith for 
examination and comment. After kindly comparing it with American species he replied that 
he was ‘unable to match the Ceylon specimen with any known American sp. and I feel that 
there is a good probability that it is not American...I think there is a good case for 
describing your plant as new.’ 

The chromosome count given for this fern by Manton (in Manton & Sledge, Joc. cit.) was 
inadvertently cited as ‘n=62’ whereas the slip accompanying the voucher specimen from 
which the count was made reads ‘n = c.60 (prob. 62)’. Dr Lovis obtained approximate counts 
from a fixing of a plant sent to Kew by Dr Faden which gave 57-59. The true number 
therefore is likely to be 58. 


V. PARATHELYPTERIS (H. [t6) Ching 
in Acta phytotax, sin. 8 : 300 p.p. (1963). Holttum in Blumea 19 : 32(1971). 


Small ferns with slender creeping rhizomes; fronds pinnate with deeply pinnatifid pinnae, 
decrescent or not; veins free, reaching the margins, costae grooved on upper surface; lower 
surface of pinnae with sessile, spherical glands and often with slender, septate hairs; sori 
indusiate, capsules eglandular and without setae; spores opaque with a narrow irregular 
wing. n= 27, 31 [c. 36]. 

About 10 species: mostly from warmer parts of mainland Asia, to New Guinea, Solomon 
and Philippine Islands, and Japan. 


1. Parathelypteris beddomei (Baker) Ching 


in Acta phytotax. sin. 8 : 302 (1963).—Nephrodium beddomei Baker in Hook. & Baker, Syn. Fil. : 267 
(1867). Type: Travancore, Beddome (?K, not seen).—Lastrea beddomei (Baker) Bedd., Ferns Brit. 
Ind. Corr.: 11 (1870); Handb. Ferns Brit. Ind. : 239 fig. 122 (1883).—Dryopteris beddomei (Baker) 
Kuntze, Rev. Gen. Pl. 11 : 812 (1891).—Thelypteris beddomei (Baker) Ching in Bull. Fan. meml. Inst. 
Biol. (Bot.) 6 : 308 (1936). Holttum, Rev. Fl. Mal. 2 : 240 (1955). 

Lastrea gracilescens sensu Bedd., Ferns South. Ind. : 38, t.110 (1863), non Aspidium gracilescens 
Blume (1828), nec Nephrodium gracilescens (Blume) Hook. (1862). Aspidium gracilescens sensu 
Thwaites, Enum. Pl. Zeyl. : 391 (1864), non Blume (1828). 


Rhizome long-creeping, sparsely scaly, fronds spaced, up to 50 cm long; stipes 5-15 cm, 
stramineous, naked or with a few scales at the base; /Jamina pinnate 20-30 x 4-6 cm, widest 
above the middle, tapering to the apex and gradually narrowed downwards, the lower pairs 
of pinnae becoming remote and much reduced in size, rhachis hairy above, minutely 
glandular-pubescent elsewhere; pinnae patent, cut almost to the costa into oblique mostly 
entire, obtuse or subacute segments, those adjacent to the rhachis often bluntly lobed, 
glabrous above save for short hairs on the costa, Jower surface bearing scattered, longish, 
white, acicular hairs on the costa and veins and sessile spherical glands; texture firm- 
herbaceous; sori submarginal, indusia small, glandular. 


Collections: Thwaites C. P. 1287 (BM; CGE; K; P; PDA). Newera Eliya, banks of streams in open places, 
Sept. 1844, Gardner 1141 (CGE). Nuwara Eliya, April 1899, Gamble 27585 (K). Same locality, 9 May 
1906, Matthew (K). Same locality, Jan. 1908, Bicknell (P). Same locality, Freeman 231, 232, 233 (BM). 


12 W. A. SLEDGE 


Same locality, Feb. 1954, Schmid 1298, 1363 (BM). Same locality at c. 1980 m, 2 Jan. 1977, Faden 77/34 
(K). Sita Eliya, patana near river, 1740 m, Oct. 1897, Pearson 226 (CGE). Between Pattipola and Horton 
Plains, marshy ground by stream, 1800 m, 20 Dec. 1950, Sledge 668 (BM). Ramboda Pass—Maturata 
track, c, 1900 m, 17 March 1954, Sledge 1315 (BM). 


In swampy places in the highest parts of the Central Province: frequent about Newara Eliya. 
Southern India, Ceylon, Sumatra, Java and widely throughout Malaysia on mountains to New 
Guinea, Taiwan, Philippines, and south Japan. 


VI. THELYPTERIS Schmidel 
in Keller, Icon. Pl.: 45, t.11, 13 (1763). 


Rhizome slender, long-creeping, growing in marshy ground; fronds pinnate with deeply 
pinnatifid pinnae, the basal ones not or little reduced; veins free, reaching the margins, 
costae grooved on upper surface, small, flat, thin scales present on lower surface of costae 
(also filamentous small ones) but sessile spherical glands absent; sori indusiate, short capitate 
hairs sometimes present on sporangia near annulus; spores spinulose. n = 35. 

Four species: north temperate Europe and Asia and south tropical and subtropical Africa, 
Madagascar, southern India, Thailand, Sumatra, New Guinea, and northern New Zealand. 


1. Thelypteris confluens (Thunb.) Morton 


in Contr. U.S. natn. Herb. 38:71 (1967). Schelpe, Fl. Zamb. Pterid.: 190, tab. 55E (1970).—Pteris 
confluens Thunb., Prodr. Pl. Cap.: 171 (1800). Type: South Africa, Thunberg (UPS). 

Aspidium thelypteris var. squamigerum Schlechtendal, Adumbr. Fil. Prom. Bon. Spei : 23, t.11 (1825). 
Type: South Africa, Cape Peninsula, Schlechtendal (HAL).—Lastrea thelypteris var. squamigera 
(Schlechtendal) Bedd., Suppl. Handb. Ferns Brit. Ind.:54 (1892).—Thelypteris palustris var. 
squamigera (Schlechtendal) Weatherby in Contr. Gray Herb. Harv. IT, 73 : 40 (1924).—Thelypteris 
squamigera (Schlechtendal) Ching in Bull. Fan meml Inst. Biol, (Bot.) 6: 329 (1936) sub 
“‘squamulosa.’ 


Lastrea fairbankii Bedd., Ferns Brit. Ind.: 254 (1867); Handb. Ferns Brit. Ind.: 240 (1883). Type: 
southern India, Pulney Mts, Beddome (K). 


Rhizome wide-creeping, black, sparsely scaly; stipes 20-30 (50) cm, stramineous, black at 
the base, glabrous; lamina 15-25 cm long, 5-8 cm wide, pinnate with deeply pinnatifid 
pinnae, broadly lanceolate or oblong in outline, basal 2-3 pairs of pinnae reduced in size and 
often more distant, rhachis glabrous or with scattered minute gland-like pubescence; pinnae 
cut almost to the costa into oblong acute or obtuse, entire lobes often with recurved margins, 
upper surface glabrous, ower surface with ovate brown scales on the costa, elsewhere 
glabrous to slightly pubescent on the midribs and veins, texture stiff, chartaceous; sori medial 
or nearer costules, indusium fimbriate at the margin. 


Collections: Swamp near Bandarawella, Uva Province, in great abundance, Sept. 1890, [no collector’s 
name] (PDA). 


Africa south of the equator, Ethiopia and Sudan, Madagascar; southern India, Ceylon, Sumatra, 
mountains of New Guinea, and New Zealand (North Island). 


The collection from Bandarawella, named Lastrea beddomeéi, is superficially similar to that 
species, but has much less markedly decrescent fronds which lack glands and white, acicular 
hairs on the under surface of the pinnae, but have ovate brown scales on the costas. The 
locality where these specimens were collected is likely to have been an area of open ground 
3-5 km west of Bandarawella, and the fern probably still grows there since modern maps still 
indicate the presence of much swamp and marsh. The elevation is 1200 m, which is the same 
as that in the Pulney Mountains locality cited by Beddome for L. fairbankii. 


VII. CYCLOSORUS Link 
Hort. Reg. Bot. Berol. 2 : 128 (1833). Holttum in Blumea 19 : 27 (1971). 


|g 


THELYPTERIDACEAE OF CEYLON 13 


Rhizome long-creeping, growing in wet ground; fronds pinnate, pinnae pinnatifid, the lower 
ones not reduced; thin, flat scales present on lower surface of costae; upper surface of costae 
grooved; basal veins anastomosing, the next pair passing to sides of sinus; lower surface of 
pinnae usually bearing acicular and/or capitate hairs and sessile, red glands; sori indusiate, 
sporangium stalk bearing multicellular gland-tipped hairs, capsules eglandular, spores 
muricate. n = 36. 

Three species: pan-tropical and subtropical. 


1. Cyclosorus interruptus (Willd.) H. It6 


in Bot. Mag., Tokyo 51 : 714 (1937) nomen tantum.—Pteris interrupta Willd. in Phytographia 1: 13, 

t.10, fig. 1 (1794). Type: southern India, Klein (B). 

Aspidium obtusatum Swartz in J. Bot. Gottingen 1800 (2): 33 (1801); Syn. Fil. : 248 (1806). Type: 
Java, Thunberg (BM). 

Aspidium goggilodus Schkuhr, Krypt. Gew. 1: 193, t.33c (1809). Type from Guyana.—Cyclosorus 
gongylodes (Schkuhr) Link, Hort. Reg. Bot. Berol. 2 : 128 (1833). Ching in Bull. Fan meml Inst. Biol. 
(Bot.) 8 : 186 (1938). Holttum, Rev. Fl. Malaya 2 : 261, fig. 148 (1955). 

Nephrodium propinquum R.Br., Prodr. Fl. Nov. Holl. : 148 (1810). Hook., Spec. Fil. 4:49 (1862). 
Bedd., Ferns South. Ind. : 32, t.89 (1863). Type: Australia, Banks (BM).—Aspidium propinquum 
(R.Br.) Thwaites, Enum. Pl. Zeyl. : 391 (1864). 

Nephrodium unitum sensu Bedd., Handb. Ferns Brit. Ind. : 268 (1883), non Polypodium unitum L. 
(1759). » 


Rhizome creeping; stipes about 50 cm long, black at the base, brown distally, glabrous save 
about the base which bears scattered brown scales when young; lamina 30-60 cm pinnate 
with up to 25 pairs of pinnae, the lowest pair not or scarcely reduced, apex of frond 
pinna-like; largest pinnae up to 15 x 1:5 cm apex acute, base truncate or broadly cuneate, cut 
down about one third into deltoid lobes with rounded sides and acute apex; veins 6-10 pairs 
per lobe, the lowest veins of adjacent lobes anastomosing forming an excurrent vein which is 
often joined by the next acroscopic vein beneath the sinus; lower surface of costae with 
scattered broad scales, the costules and veins bearing sessile, spherical, reddish or orange 
glands, otherwise both surfaces glabrous (in Ceylon plants) save for some ciliate hairs on 
the margins; textures chartaceous; sori medial on the veins, absent from the lowest pair, 
indusia glabrous or with a few hairs. 


Collections: Thwaites C.P. 705 (BM; CGE; K; P; PDA). Colombo, coll. Randall 1870, 1871, Rawson 
3245 (BM). Caltura, Macrae 217 (CGE; E; K; P). Galle, Freeman 262 (BM). Tamankaduwa, 1893, 
Nevill (PDA). Near Ratnapura, marshy ground by road, 3 Jan. 1951, Sledge 798 (BM). Marsh near 
Bentota, Western Province, 19 Jan. 1951, Sledge 880 (BM). Unlocalised: Barkly 14, 16 (BM). 1839, 
Mackenzie (K). Skinner (K). Ferguson (US). 
Open, marshy places at low elevations in western and southern parts of Ceylon. 

Tropical and subtropical Africa, Madagascar; India eastwards to China, Japan and the 
Philippines, and south through Malesia to New Zealand; Polynesia; tropical America 
including Brazil. 


An exclusively low altitude fern; the only other thelypteroid ferns in Ceylon which are 
also confined to low elevations are Ampelopteris prolifera, Trigonospora glandulosa and T. 
zeylanica. 

The under surfaces of the pinnae seem always to lack hairs in Ceylonese specimens. In 
other parts of its very wide range they vary from glabrous to densely villous. Mettenius 
created the varieties glabrum and hirsutum to cover such variation and Ching has upheld 
them. Others have not done so because there appears to be a continuum between the two 
extremes. 

VII. AMPELOPTERIS Kunze 
Bot. Ztg. 6: 114 (1848). 


Rhizome wide-creeping; fronds of indefinite apical growth proliferating from buds borne 


14 W. A. SLEDGE 


distally on the rhachis; forked unicellular hairs present on rhachis; pinnae subentire, most 
veins from adjacent costules anastomosing; sori exindusiate, round or somewhat elongate, 
hairs on stalks of sporangia bearing a terminal glandular cell. n = 36. 

A monotypic genus distributed throughout the wetter parts of tropical and subtropical 
regions of the Old World. 


1. Ampelopteris prolifera (Retz.) Copel. 


Gen. Fil.: 144 (1947).—Hemionitis prolifera Retz., Obs. Bot. 6:36 (1791). Type: southern India 
Koenig (GOET)—Meniscium proliferum (Retz.) Swartz, Syn. Fil.: 19. 207 (1806). K. Iwats. in 
Mem. Coll. Sci. Kyoto Univ. B, 31: 196 (1965).—Goniopteris prolifera (Retz.) C. Presl, Tent. 
Pterid. : 183 (1836). Bedd., Ferns South. Ind. : t.172 (1864); Handb. Ferns Brit. Ind. : 296, fig. 153 
(1883).—Dryopteris prolifera (Retz.) Christensen, Jndex Fil. : 286 (1905). 

Ampelopteris elegans Kunze in Bot. Ztg. 6: 114 (1848). Origin unknown. 


Rhizome creeping, sparsely scaly, stipes close together; fronds pinnate some pinnae bearing 
axillary buds which may remain dormant or may form tufts of secondary fronds; fronds of 
two kinds, some 30-60 cm long, imparipinnate with up to 10 pairs of pinnae, others of 
indefinite apical growth, arching over and rooting at or near the tip, their distal pinnae much 
shortened; pinnae up to 15cm long, 1-1:75 cm broad, subsessile with truncate base and 
acute apex, becoming obtuse in the shortened distal pinnae, margins crenate or subentire; 
veins up to 10 pairs, 4-6 pairs uniting to form a sinuous excurrent vein; rhachis bearing 
some unicellular forked hairs; costa hairy above and with a few ciliate scales beneath, veins 
with scattered hairs on lower surface; texture firm-herbaceous; sori oblong or elongate, 
becoming confluent in age, exindusiate. 


Collections: Thwaites C.P. 3916 (PDA). Tamankaduwa District, near Matale boundary, 1893, Nevill 
(PDA). Minnery, Beckett 1131 (PDA). Uma-Oya, in water, 1881, [no collector’s name] (PDA). Kannia, 
hot wells, Dec. 1885, W. Fferguson] (PDA). Hot wells close to Trincomalee, Freeman 295, 296 (BM). 
Kannia Hot wells near Trincomalee, in marsh, 16 Jan. 1951, Sledge 879 (BM). 
Rare in Ceylon, all localised records are on the eastern side of the central massif at low 
elevations. 

Widespread in the tropics of the Old World from Africa to New Caledonia. 


A marsh species growing in open places or in light shade. Those fronds which have 
indefinite apical growth produce shortened pinnae and such fronds always seem to remain 
sterile. 


IX. TRIGONOSPORA Holttum 


in Blumea 19 : 29 (1971). 


Caudex erect; basal pinnae not or little shortened and never reduced to auricles on the stipe, 
pinnae usually deeply lobed; veins free the basal acroscopic one running to the base of the 
sinus between two segments, the basal basiscopic one reaching the edge above the base of the 
sinus; sori indusiate; spores trilete, minutely papillose. n = 36. 

Trigonospora is unique amongst thelypteroid ferns in having trilete spores. All species 
commonly grow amongst stones in or close by river beds. The genus is confined to south- 
eastern Asia and was estimated by Holttum (in Blumea Joc. cit.) to contain about eight 
species with the probability of some other species awaiting description. Over most of its area 
of distribution the species are clearly defined (and often very uniform), but in southern India 
and Ceylon, the degree of variability reaches its maximum and the species are, in 
consequence, more difficult to distinguish. This diversity led Hooker and Thwaites to 
include all plants from there within one variable species which they identified as Blume’s 
Javan plant Aspidium calcaratum. Beddome recognised a number of different variants which 
he treated as varieties under Lastrea calcarata (Blume) Moore. 


THELYPTERIDACEAE OF CEYLON 15 


I believe that most Ceylon specimens can be referred readily to one or other of the species 
enumerated below. Of these species, two are widely distributed in Ceylon, and the great 
majority of specimens in herbaria are referable to them. Trigonospora glandulosa is very 
distinct but is so far only known from a single gathering; 7. angustifrons, which is also rare, is 
readily distinguished from T. calcarata but, nevertheless, its relation to that species requires 
further investigation; the identity of specimens referred to T. ciliata (Benth.) Holttum is more 
problematical. 

Incomplete specimens, such as isolated fertile fronds, are sometimes difficult to name with 
confidence, and a few gatherings appear to be intermediates with mixed characters. Whether 
these are indicative of introgression can only be decided by experimental work. One of my 
gatherings (Sledge 752) differs from all others in having numerous forked veins with 
occasional anastomoses occurring between them in the pinna segments of the sterile frond; 
also present are a few small subsessile, colourless, clubbed hairs or glands on the veins 
beneath. These characters are most readily accounted for as a consequence of hybrid origin, 
the other parent being presumably Christella parasitica. No spores are present on the fertile 
frond but this may be due to age. Two gatherings in the British Museum, one of them 
collected by Moon in Ceylon, are also remarkable for the numerous proliferating buds in the 
axils of the pinnae. Further cytotaxonomic and cultural, as well as field observations are 
required before definitive conclusions as to the relationships can be established. 


Key to species of Trigonospora 


l Base of pinnae truncate with pinnalobesclosetorhachis . . . . .. . 2 
Base of pinnae cuneate with no pinna lobes close to rhachis 7. T. zeylanica (p. 24) 

2(1) | Lower surface of pinnae & indusia eglandular ‘ 2 ' : 3 
Lower surface of pinnae & indusia glandular ig , 6. 7, glandulosa (p. 23) 


3(2) Basal acroscopic lobes free at least in lower pinnae, lobes very oblique, indusia glabrous 4 
Basal acroscopic lobes not free, lobes oblique, indusia hairy or glabrous toate 

4(3) | Lamina 2-3 times as long as wide, pinnae oblong, hairy at least on the costae . 4.T. 

calcarata (p. 21) 

= Lamina 3-6 times as long as wide, pinnae fusiform, quiteglabrous .  .. py be 

angustifrons (p. 23) 

5(3)  Pinnae 4 cm or less, cut § way to costa, tips blunt or acute 2A oe 3. T. ciliata (p. 20) 

- Pinnae 4 cm or more, cut 3 or more to costa, tips acuminate or caudate Sia 2 5 6 
6(5)  Pinnae 4-8 cmlongwithabout 10 pairsoflobes . . . 2. T. obtusiloba (p. 18) 
- Pinnae 8-13 cm long with 15-20 pairsoflobes_ . a |. T. caudipinna (p. 15) 


nN 


1. Trigonospora caudipinna (Ching) Sledge, comb. nov. (Fig. 2A) 


Thelypteris caudipinna Ching in Bull. Fan meml Inst. Biol. (Bot.) 6: 288 (1936). Type: Hainan, 
Hancock 108 (K).—Pseudocyclosorus caudipinnus (Ching) Ching in Acta phytotax. sin. 8 : 324 (1963). 

Aspidium ciliatum Wall. Numer. List. : 351 (1829) nom. nud.—Nephrodium ciliatum C. B. Clarke in 
Trans. Linn. Soc. (Bot.) Il, 1: 514 (1880) p.p., nom., illeg., excl. Lastrea sericea sensu Bedd.— 
Thelypteris ciliata sensu Ching in Bull. Fan mem Inst. Biol. (Bot.) 6 : 289 (1936) p.p., excl. Aspidium 
ciliatum Wall. ex Benth. and Lastrea sericea sensu Bedd.—Pseudocyclosorus ciliatus sensu Ching in 
Acta phytotax. sin. 8 : 324 (1963) p.p. 

Aspidium canum Wall. Numer. List : 387 (1829) nom. nud.—Lastrea cana J.Sm., Cat. Cult. Ferns : 57 
(1857) p.p., nom. nud.—Thelypteris cana Ching in Bull. Fan mem Inst. Biol. (Bot.) 6 : 291 (1936) 


p.p. 

Lastrea falciloba Bedd., Ferns S. Ind. :37, t.105 (1863), non Nephrodium falcilobum Hook. Type: 
Anamallays, 3000’, Beddome (K). 

Lastrea bergiana sensu Bedd., Ferns S. Ind. & Brit. Ind. suppl.: 16 t.370 (1876), non Polypodium 
bergianum Schlechter. 

Lastrea calcarata var. ciliata Bedd., Handb. Ferns Brit. Ind. : 235, fig. 121 (1883) 

Trigonospora ciliata var. angustiloba Holttum in Saldanha & Nicolson, Fi. Hassan Dist. Karnataka, 
India : 866 (1976). Type from southern India. 


16 W. A. SLEDGE 


Fig. 2 Pinnae: A—Trigonospora caudipinna (Ching) Sledge (Sledge 1341), x 1; B—T. ciliata (Wall. 
ex Benth.) Holttum (Sledge 710), fertile and sterile, x 1:5; C-T. zeylanica (Ching) Sledge (Sledge 
812), x 1:5. 


Caudex erect, fronds up to 1 m long. Lamina 30-40 (-60) x 15-20 (—25) cm, broadly oblong 
to ovate-oblong with 15-20 pairs of free, sessile pinnae, the lowest pair usually deflexed and 
somewhat narrowed at the base at least on the basiscopic side. Pinnae pinnatifid to 2/3 or 
more, 8-13 x 15-2 (2°5)cm with 15 or more pairs of rounded lobes 34 (5) mm wide, the 
basal acroscopic lobes sometimes more deeply cleft but not, or very rarely, free to the base, 
often enlarged with forked veins and usually incurved close to or underlying the rhachis, 
apex of pinnae with attenuate or caudate, entire tips 1-2 (3) cm long; costa, costules and veins 
usually sparsely hairy beneath, margins with scattered hairs, surfaces glabrous or nearly so. 
Indusia glabrous or hairy. 


Collections: Thwaites C.P. 1363 (CGE; K; P; PDA). Rambodde, shady banks, June 1845, Gardner 
1107 (BM; CGE; K). Gardner 49 (P). Hantane, 1868, Robinson 156 (K). Yelumali, Namunukula, 12 
March 1907, J. M. Silva (PDA). Mawanella, Sab. Province, 27 Jan. 1954, Schmid 1070 (BM). 
Ramboda Pass—Maturata track, c. 1900 m, 17 March 1954, Sledge 1303 (BM). Hakgala, edge of forest, 
1800 m, 20 March 1954, Sledge 1341 (BM). 


Moist banks especially by streams mainly in mountain forests. 
South India, Ceylon, Nepal, Sikkim, Assam, Burma and China (Hainan Dao). 


THELYPTERIDACEAE OF CEYLON 17 


The type sheet of Thelypteris caudipinna Ching at Kew carries two detached fronds. These 
are 56 and 39 cm long, with laminas 28 and 23 cm long respectively; the pinnae are up to 
9 x 1:5 cm long, with rather broad oblique lobes and with caudate apices 2-3 cm in length on 
the larger frond, but not exceeding 1-2 cm on the second frond. The basal pair of pinnae is 
deflexed, and the basal acroscopic segments of the pinnae are slightly enlarged and incurved 
to the rhachis. The indusia are sparsely hairy on one frond, but glabrous or with only 
occasional hairs on the other. 

Apart from the exaggeratedly caudate tips to the pinnae on one of these fronds, the 
specimens match many gatherings from northern and southern India and Ceylon. Comparable 
specimens from India and Ceylon all have attenuate to caudate, entire tips to the 
pinnae, not infrequently reaching 2 cm in length. Since the presence of even longer caudate 
apices in one only of the type specimens is not consistently linked with any other characters 
which cannot be found in many other gatherings, this feature is evidently no more than an 
extreme development in an individual specimen, and cannot be regarded as of specific 
significance. Further evidence that Hancock’s gathering from Hainan represents an abnor- 
mally developed specimen is provided by the fact that it has remained the only gathering in 
the Thelypteris caudipinna folder at Kew and no other collection is known to exist. 


John Smith’s Lastrea cana, later validated by Ching as Thelypteris cana, is the same 
species, as judged by three sheets in Smith’s herbarium at the British Museum. Each 
sheet carries three detached fronds. Holttum has identified a frond on one sheet as 
Amauropelta bergiana (Schlechtendal) Holttum, but the other eight are all fertile fronds ofa 
large species of Trigonospora, mostly with about 20 pairs of pinnae. These specimens match 
the fronds in Wallich’s collection named Aspidium ciliatum (Wall. 351) and A. canum 
(Wall. 387), and they agree much more closely with Hancock’s Hainan fern and Beddome’s 
specimen from southern India than with Bowring’s type specimen of 7rigonospora ciliata 
(Benth.) Holttum from Hong Kong. The latter is a small fern with fronds 17°5-22°5 cm in 
length, the pinnae are without caudate or acuminate tips and the largest measure 4°5 x | cm. 
Part of the confusion which has prevailed between 7. caudipinna and T. ciliata can be 
ascribed to the inexplicable statement in Hooker’s J. Bot. 9 : 338 (1857) that the Hong Kong 
plant ‘entirely agrees’ with Wallich’s Himalayan plants, when in fact they are notably different. 
This misleading statement was accepted by Ching (/oc. cit. : 290) whose T. ciliata is a mixture of 
the two species. 

I have been unable to discover on what grounds Ching stated that the type of The/lypteris 
cana was a gathering by Wight from south India. Smith cited no specimens, gave the 
provenance simply as East Indies, and none of his specimens at the British Museum were 
gathered by Wight. Ching implied that 7. cana was a south Indian species but this is refuted 
by one of John Smith’s sheets carrying specimens from Nepal, whilst both the synonyms cited by 
Ching were based on gatherings from north India. The three sheets named Lastrea cana by 
Smith are from: (a) Nepal: Wallich and Hamilton, (b) Madras: Hook. fil. & Thomson 249, and (c) 
Ceylon: Gardner 1107. The second of these sheets, the one on which all three fronds are clearly 
conspecific (though none has the ‘densely hirsute’ indusia indicated by Ching), is here chosen as 
the lectotype. 


Beddome’s specimen from the Anamallay Hills, from which was prepared his illustration 
of Lastrea falciloba (later, in his Handbook, named L. calcarata var. ciliata), is the same as 
the Hainan plant. It has a lamina 48 x 20 cm, pinnae up to 10 x 2 cm with markedly caudate 
apices and 15 or more pairs of pinna segments, 3 mm wide. The indusia are hairy. Ching (/oc. 
cit. : 298) was incorrect in referring Beddome’s illustration to Thelypteris falciloba (Hook.) 
Ching (later renamed Pseudocyclosorus falcilobus (Hook.) Ching), for Beddome specifically 
refers in the description in his Handbook to the absence on the stipe, in contrast to 
his var. falciloba which is ‘furnished with auricles below the frond.’ Large plants, closely 
similar to Beddome’s specimen, occur in Sikkim, Assam and Burma and doubtless over the 
intervening region to Hainan. In Ceylon they are less common than the smaller Trigono- 
spora obtusiloba, and localised specimens mostly come from high altitudes. Robinson 156 


18 W. A. SLEDGE 


(K) closely matches the type of T. caudipinna, differing only in the shorter caudate tips to the 
pinnae. 

Clarke’s specimens of Nephrodium ciliatum and his description of the common form in 
Khasia as having caudate pinnae leave no doubt that Trigonospora caudipinna was the 
species intended. He correctly treated Lastrea bergiana sensu Bedd. as the same species. 
Although this is without reduced pinnae or auricles on the stipe, in his Handbook, Beddome 
later equated the species with Baker’s L. cana in which the fronds are decrescent with 
auricled stipes. Beddome’s own illustration of L. cana (in Ferns Brit. Ind.: t.307) depicts 
these characters, and is markedly different from his illustration of L. bergiana, although both 
illustrations are cited in the Handbook (p. 239) as portraying the same species. Clarke 
included L. sericea Scott ex Bedd. as a synonym of his Nephrodium ciliatum. This is similar 
in stature to Trigonospora caudipinna from which it differs in its short, wide and blunt 
pinnae. In other respects it does not vary significantly from 7. caudipinna, to which it is 
certainly closely related. It should perhaps be regarded as a variant; however, such a treat- 
ment would necessitate the substitution of the earlier sericea for caudipinna as the specific 
epithet of the collective species. 

Holttum’s Trigonospora ciliata var. angustiloba, described as having ‘frondibus multo 
majoribus, pinnis acuminatis’ and identified with Beddome’s illustration in Ferns S. India 
and the Handbook, is also T. caudipinna. The reason for the author’s choice of epithet is not 
clear as there is no reference to the lobes of the pinnae being narrow in the type description. 


2. Trigonospora obtusiloba Sledge, sp. nov. (Fig. 3) 


Caudex erectus. Frondes usque ad 50 cm longae, fertiles stipitibus longis. Lamina in ambitu 
oblonga, ovata-oblonga vel deltoidea-ovata, plerumque cum 7-10 sed usque ad 15 paribus 
pinnarum liberarum sessilium, infimas deflexes basi contractas. Pinnae steriles (3) 4-6 
(8) x 1-2 cm pinnatifidae ad 2/3, circa 10 paribus segmentorum subpatentium vel parum 
obliquorum, latorum, obtuse rotundatorum usque ad 5 mm latorum; frondes fertiles pinnis 
segmentisque angustioribus, segmenta basalia acroscopica plerumque aliquantum aucta, 
incurvata et saepe subter rhachides inflexa sed interdum vix ceteris dissimilia; pinna apice 
integra, acuta vel acuminata; costae, costulaeque venae subter pilis numeris variis, paginae 
subglabrae. Indusia plerumque hirsuta, interdum sparsissima vel glabra. 


Caudex erect; fronds up to 50cm, the fertile ones with long stipes. Lamina oblong, ovate- 
oblong or deltoid-ovate in outline, commonly with 7-10 but up to 15 pairs of free sessile 
pinnae, the lowest pair deflexed and contracted at the base. Sterile pinnae (3) 4-6 
(9) x 1-2 cm, pinnatifid to 2/3 with about 10 pairs of subpatent or slightly oblique, broad and 
bluntly rounded segments up to 5mm wide; fertile fronds with narrower pinnae and 
segments; basal acroscopic segments usually a little enlarged, incurved and often underlying 
the rhachis but sometimes scarcely different from the rest; pinnae apex entire, acute or 
acuminate; costa, costules and veins with variable amounts of hair below, surfaces 
subglabrous. Indusia usually hairy, sometimes very sparsely so or glabrous. 


Collections: Adam’s Peak, 14 Feb. 1908, Matthew (K; P). Same locality, at 1800 m, 14 Dec. 1950, 
Sledge 609 (BM). Nuwara Eliya, 1800 m, 10 May 1906, Matthew 622 (K). Same locality at 1980 m, 2 
Jan. 1977, Faden 77/181 (K). Kuda Oya on Ramboda road, 1650-1700 m, 28 Dec. 1950, Ballard 1291 
(K). Hakgala, 1800-1900 m, 27 Dec. 1950, Sledge 754 (BM). Sudagalla, 1050 m, 2 Feb. 1891, Hancock 
22 (K). Riverstone Estate, Matale District at 1100 m, 19 Jan. 1977, Faden 77/181 (K). Ratnapura 
District, near stream, 750 m, 16 Nov. 1976, Faden 76/309 (K). Pahale Hewessa, near stream in jungle, 
45 m, 20 Jan. 1951, Ballard 1527 (K); Sledge 889 (BM). Sinha Raja Forest near Hedigala, 75 m, 5 Jan. 
1951, Ballard 1393 (K); Sledge 814 (BM, holotype). Sinha Raja Forest above Beverley Estate, 
Deniyaya, by stream 900m, 12 March 1954, Sledge 1287 (BM). Kottawa Forest Reserve, 21 Jan. 1951, 
Ballard 1536 (K). Same locality, 60 m, 1 April 1954, Sledge 1378 (BM), Near Udagama near stream in 
jungle, 90 m, 21 Jan. 1951, Sledge 907 (BM). Kotmalee, 1847, Fortescue (CGE). Unlocalised: ‘Ceylon 4000 
ft.’, Beddome (K). Wall spec. in herb. Blanford ex herb. C. W. Hope (P). Walker (K). Gardner 1363 (K). 
Gardner 1107 partim (CGE). 


THELYPTERIDACEAE OF CEYLON 19 


‘aw Ry 
Aa), 
Aw hy Ab 
ee 
SJ AUN 

0 ZA NN 
Fig: Ae 
i 


Fig. 3 Trigonospora obtusiloba Sledge: plant (Sledge 814), x 0-4. 


By rivers and streams on shaded ground at all elevations up to 2000 m. 
Known only from Ceylon. 


This is the commonest species of Trigonospora in Ceylon. It is characterised by the marked 
distinction between sterile and fertile fronds, the latter having conspicuously longer stipes, 
often twice the length of the lamina and overtopping the sterile fronds. The sterile fronds 
have broad pinnae the segments of which are wide and rounded, the basal pair being deflexed 
and contracted at the base, at least on the basiscopic side. The fertile fronds have much 
narrower pinnae and pinna-segments, but fronds with broad pinnae similar to those of the 


20 W. A. SLEDGE 


sterile fronds with the segments bearing scattered sori are also produced; such fronds rarely, if 
ever, produce sori on each of the paired veins of the pinna segments, as in typical, fully fertile 
fronds. Sometimes the tips of the pinnae are long-attenuate or even caudate as in TJ. 
caudipinna, though T. obtusiloba is always a smaller species. The narrower fertile pinnae, 
and hence reduced size of the pinna segments in 7. obtusiloba, results in the segments not 
being broadly rounded as in the sterile fonds; large detached fertile fronds of T. obtusiloba 
are therefore not easily distinguished from small fertile fronds of T. caudipinna. One of the 
difficulties in naming the species of Trigonospora lies in the fact that many gatherings consist 
only of detached fertile fronds, thus the valuable information afforded by comparison with 
sterile fronds from the same plant is not available. 

Trigonospora obtusiloba and T. caudipinna have not previously been distinguished from 
one another. Both are usually labelled T. ciliata or T. calcarata var. ciliata in herbaria. Although 
Beddome’s Lastrea calcarata var. ciliata was based on a large plant—oniginally named L. 
falciloba—which.I consider to be identical with Ching’s Thelypteris caudipinna, he later failed to 
distinguish between this and T. obtusiloba for there is a gathering of the latter from Ceylon 
named L. calcarata var. ciliata by him at Kew. 

Trigonospora obtusiloba bridges the gap in size between T. caudipinna and T. ciliata, and 
this has doubtless contributed to the confusion between the two species. In T. ciliata the 
pinnae are less deeply lobed, the lobes of the sterile pinnae are never so broad and blunt, and 
the tips of the pinnae are usually acute, sometimes shortly acuminate, but never caudate. 


3. Trigonospora ciliata (Wall. ex Benth.) Holttum (Fig. 2B) 


in Blumea 19 : 29 (1971).—Lastrea ciliata Hook. in Hooker’s J. Bot. 9 : 338 (1857), nom. illeg., non L. 
ciliata Liebm. (1849).—Aspidium ciliatum Wall. ex Benth., Fl. Hong Kong: 455 (1861). Type: Hong 
Kong, Bowring 25 (K).—Dryopteris ciliata (Wall. ex Benth.) Christensen ex Wu, Wong & Pong in 
Bull. Dept. Biol. Coll. Sci. Sun Yatsen Univ. 3 : 30, pl. 6 (1932).—Thelypteris ciliata (Wall. ex Benth.) 
Ching in Bull. Fan meml Inst. Biol. (Bot.) 6 : 289 (1936) p.p. excl. Aspidium ciliatum Wall. nom. 
nud., Lastrea sericea sensu Bedd., and Nephrodium ciliatum C. B. Clarke nom. illeg— 
Pseudocyclosorus ciliatus (Wall. ex Benth.) Ching in Acta phytotax. sin, 8 : 324 (1963) p.p. 

Dryopteris pseudocalcarata Christensen, Index Fil. Suppl. 3:95 (1934) p.p., excl. Aspidium ciliatum 
Wall. nom. nud., Lastrea sericea sensu Bedd., and Nephrodium ciliatum C. B. Clarke nom. illeg. 


Caudex erect; fronds up to 40 cm, the fertile ones with long stipes. Lamina oblong with up to 
15 pairs of free pinnae, the lowest pair usually not deflexed. Sterile pinnae about 4 x 1 cm, 
pinnatifid half way to the costa with about 10-12 pairs of oblique, falcate segments, the basal 
acroscopic ones enlarged and incurved; apex of pinnae entire, acute or acuminate, costa, 
costules and veins hairy. Indusium (in Ceylonese specimens) naked or with sparse hairs. 


Collections: Pas Dun Corle, Aug. 1865, Thwaites C.P. 992 (P). Between Hakgala and Nuwara Eliya, 
moist ground by track through jungle, 1650 m, 23 Dec. 1950, Sledge 710 (BM), Holttum 39161 (SING). 


Very rare, or possibly overlooked; the absence of specimens in herbaria suggests the former. 
South China and Thailand westwards to north-east India and southwards to northern 
Malaysia, Sumatra and Ceylon. 


The type of Trigonospora ciliata, which is from Hong Kong, is considered by Holttum (see 
Reinwardtia 8 : 503-507 (1974)) to agree well with Malayan material. North India was also 
included in its distribution by both Holttum and Ching, although I have seen no specimen 
from this area which matches the Hong Kong or Malayan plants. 

The first two collections cited above are not closely similar but are best referred to 
Trigonospora ciliata. Neither could be included in any other Ceylonese species. My 
own gathering (no. 7/0) closely resembles the Hong Kong type in size and form. It differs, 
however, as does Thwaites’ specimen of C.P. 992 at Paris (which is manifestly different from 
other gatherings bearing the same number), in having indusia which are either glabrous or 
only have very few hairs; in the type the indusia have abundant, long, setose hairs. However, 


——t 


THELYPTERIDACEAE OF CEYLON 21 


not all specimens referred to this species are similarly clothed; Holttum (loc. cit.) described 
the indusia as ‘very hairy’ but in his description of Malayan plants they are alluded to as only 
‘ciliate.’ There are considerable differences also in the degree of hairyness and length of hairs 
on the fronds of specimens referred to T. ciliata. Since plants of T. caudipinna may have 
densely hairy, sparsely hairy, or glabrous indusia, I do not believe that the presence or 
absence of hairs on the indusia is of significant taxonomic importance in the species of 
Trigonospora. Even in T. calcarata, in which the indusia are nearly always glabrous, I have 
seen specimens in which some hairs are present on the indusia. 

The plants recorded as Thelypteris ciliata (Wall.) Ching by Manton & Sledge (Phil. Trans. 
Roy, Soc. Lond. B, 238 : 137, 1954) were probably Trigonospora obtusiloba. 


4. Trigonospora calcarata (Blume) Holttum (Fig. 4B) 


in Reinwardtia 8 : 506 (1974). Aspidium calcaratum Blume, Enum. Pl. Jay. : 159 (1828). Type: Java, 
Blume (L).—Lastrea calcarata (Blume) Moore, Jnd. Fil.: 87 (1858)—Nephrodium calcaratum 
(Blume) Hook., Spec. Fil. 4:93 (1862) p.p.—Dryopteris calcarata (Blume) Kuntze, Rey. Gen. Pl. 
2: 812 (1891).—Thelypteris calcarata (Blume) Ching in Bull. Fan meml Inst. Biol. (Bot.) 6: 288 
(1936). 


Caudex erect. Fronds up to 40 cm long, stipes hairy or glabrescent save in the groove, those 
of the fertile fronds longer than those of the sterile ones. Lamina oblong or ovate-oblong, up 
to 12 cm wide and 2-3 times as long with 10-13 pairs of free pinnae below the pinnatifid 
apex, the lowermost pair usually not deflexed. Pinnae oblong 4-7 x 0-75-1:0 cm, rarely 
larger, pinnatifid from 3 to } to the costa with up to 12 pairs of narrow, very oblique, falcate 
segments, 1°5-3 mm. wide, the basal acroscopic segments at least in the lower pinnae quite 
free to the base (or rarely very nearly so), erect and often a little elongated and lying close to 
the rhachis, the second pair of segments also sometimes free to the base, these and the 
remaining segments very oblique, pinna apex acute or shortly attenuate, entire; surfaces 
glabrous above save on the costa, subglabrous or with scattered rather long hairs on the costa 
and veins below. Indusia glabrous. 


Collections: Thwaites C.P. 1363 (P, partim; PDA, partim). Thwaites C.P. 3273 (BM; CGE; K; P; PDA, 
partim). Moist woods at the foot of Adam’s Peak, March 1846, Gardner 1250 (CGE; K). Foot of 
Adam’s Peak, Carney near Ratnapura 240 m, 9 March 1954, Sledge 1246A (BM). Kitulgala, amongst 
rocks by riverside [low alt.], 28 Aug. 1927, Alston 897 (PDA). Deniyaya, Southern Province, 550 m, 5 
Feb. 1954, Schmid 1138 (BM). 3-4 miles east of Madugoda, Central Province, amongst stones in 
stream in jungle, 750 m, 8 Jan. 1954, Sledge 937 (BM). Hunnasgiriya, Central Province, in stream bed 
amongst rocks in jungle, 870 m, 16 Jan. 1954, Sledge 981 (BM). Gallebodde, by stream in jungle 
600 m, 26 Jan. 1954, Sledge 1042 (BM). 


Stony ground by or in river beds in shade at low or moderate elevations. 
Sumatra, Java and Ceylon. 


The distinctive characters of Trigonospora calcarata are the deeply divided pinnae with very 
oblique segments, the basal segments normally (and sometimes the second pair also) being 
quite free down to the costa and often having a small lobe at the base. The indusia are almost 
always glabrous; I have seen two gatherings with hairs on the indusia but they were otherwise 
indistinguishable from plants with glabrous indusia. The segments of the fertile pinnae are 
narrow, and the sterile pinnae do not differ much from the fertile ones except that the pinna 
segments tend to be less narrow and the lobing less deep, sometimes to little more than half 
way to the costa, though they are still strongly forwardly-directed. 

Ching did not question Ceylon and Jayan plants as being conspecific. Holttum (in 
Reinwardtia, loc. cit.) queried the occurrence of true Trigonospora calcarata in Ceylon on 
the grounds that plants from there, though similar to Javan specimens, were considerably 
larger. Javan specimens are very uniform. Blume’s type has one frond 35 cm long with 
pinnae 2-3 cm long but of the 14 Javan sheets examined, most have shorter fronds and all 
have pinnae 2-3 cm long with 4 cm as the longest. Specimens from Ceylon are mostly both 


22 W. A. SLEDGE 


DEE 
i? 
—- 


Or 
R= 


Fig. 4 A-Trigonospora angustifrons Sledge, frond (Sledge 1246, holotype), x 0°5; B—T. calcarata 
(Blume) Holttum, pinna (Sledge 1042),x2; C-T. glandulosa Sledge, pinna (Sledge 808, 
holotype), x 2; D-part of under surface of C, showing subsessile glands on lamina surface and 
indusium, x 30. 


considerably larger and more variable. The average length of the pinnae is twice that of 
Javan plants. Though such plants appear to be different they display all the essential 
characters which distinguish T. calcarata from other species of Trigonospora without having 
any additional distinctions, and since some gatherings from Ceylon are scarcely 
distinguishable in size, as well as in frond architecture, from Javan plants, I believe all should 
be included within T. calcarata. 


THELYPTERIDACEAE OF CEYLON 23 


Thwaites was probably influenced by Hooker when he included all Trigonosporas within 
the one species, Aspidium calcaratum, for he was evidently aware of the difference between 
Trigonospora calcarata s.str. and T. caudipinna since, of the two Thwaites C.P. numbers 
quoted in his Enumeratio, specimens sent out as C.P. 1363 were nearly all 7. caudipinna, whilst 
C.P. 3273 was nearly all T. calcarata. Gardner’s two numbered but unnamed collections 
were similarly not mixtures of two species; his no. 1107 was invariably T. caudipinna, whilst 
no. 1250 was consistently 7. calcarata. 


5. Trigonospora angustifrons Sledge, sp. nov. (Fig. 4A) 


Caudex erectus. Frondes 20-40 (45) cm longae, stipites breves, fertiles quam steriles vix 
longiores. Stipes rhachisque glabra praeter in sulco dorsali rhachidis pubescentes. Lamina 
linearis-oblonga 5-6 cm lata, triplo ad sexies longior cum fere 15 jugis pinnarum liberarum 
ascendentium, infima juga non deflexa. Pinnae fusiformis usque ad 4 cm longae 5-8 mm 
latae, cum octo vel paucioribus jugis segmentorum obliquorum falcatorum, fere ad 
dimidium costae (raro ultra) incisae in parte medio latissimo, basin versus attenuatae 
paribus infimis segmentorum omnino liberis, erectis, acroscopicis saepe leviter lobatis et 
rhachidi postice applicatis; pinnae distales serratae vel crenatae supremae integrae; apices 
pinnarum acuminati integri; paginae supra et infra omnino glabrae, margines non ciliati. 
Indusia glabra. 


Caudex erect. Fronds 20-40 (45) cm long, stipes short, the fertile ones not much longer than 
those of the sterile fronds. Stipe and rhachis glabrous beneath, hairy in the dorsal groove of 
the rhachis. Lamina linear-oblong, 5-6 cm wide and 3-6 times as long, with about 15 pairs 
of free ascending pinnae, the lowest pair not deflexed. Pinnae fusiform, up to 4 cm long, 
5-8 mm wide with 8 pairs or fewer of oblique falcate segments, cut half way (rarely more) to 
the costa in the broadest middle region, narrowed at the base with the lowest pair of segments 
quite free, the acroscopic segment often slightly lobed and closely applied to the rhachis; 
distal pinnae serrate or crenate, the uppermost entire; apex of pinnae acuminate, entire; 
surfaces quite glabrous above and beneath or with a few hairs on the costa above; margins 
not ciliate. Indusia glabrous. 


Collections: At foot of Adam’s Peak near Carney, Ratnapura District, by river, 240 m, 9 March 1954, 
Sledge 1246 (BM, holotype). Near Rassagala east of Ratnapura, edge of river in forest, 750 m, 16 Nov. 1976, 
Faden 76/316 (K). Adam’s Peak, Moon (BM). Unlocalised: F.D’A Vincent (CGE). 


By streams and rivers at low to medium altitudes about Adam’s Peak. 
Endemic to Ceylon. 


A well-marked species related to Trigonospora calcarata, but distinguished by its narrow 
fronds, often six times as long as broad, and by the whole plant, save the dorsal groove of the 
stipe and rhachis, being quite glabrous. The pinnae resemble those of 7. zeylanica in size and 
their fusiform outline with strongly cuneate bases, but in 7. angustifrons a pair of free, erect 
basal segments adjacent to the rhachis is always present. The next pair of segments appear to 
be somewhat distant from the basal pair on account of their being very oblique to the pinna 
axis with cuneate hinder margins. 

Apparently a rare species; the three localised gatherings all come from the Peak 
Wilderness area. 


6. Trigonospora glandulosa Sledge, sp. nov. (Fig. 4 C, D) 


Caudex erectus. Frondes usque ad 40 cm longae, fertiles stipitibus quam steriles longioribus 
non prominenter, stipes rhachisque omnino cinerei-tomentosi cum pilis brevibus crispatis et 
(praesertim in rhachidi) pilis longioribus patulis acicularibus | mm longis vestitis. Lamina 
15-30 x 6-8 cm in ambitu anguste oblonga vel elliptica sursum angustata, fere 10-12 
interdum usque ad 16 juga pinnarum liberarum infimis non deflexas. Pinnae 3-5 x 0:5- 


24 W. A. SLEDGE 


1:0 cm profunde pinnatifidae usque ad 10 paria segmentorum obliquorum oblongorum 
obtusorum falcatorum 2 mm latorum, segmenta infima omnino libera, interdum breviter 
petiolata saepe prope basin lobata leniter ampliata contigua et saepe subter rhachin 
disposita; aliquando jugum secundum segmentorum similiter ad basin liberum, residuum 
pinnae paulatim minus profunde partitum in segmentis obliquis obtusis falcatis; pinnae ad 
apicem obtusae vel acutae pinnae supremae et illae ad partem apicis adnatae omnino 
integrae. Costae costulaeque utrinque pilis acicularibus instructae et margines segmentorum 
pilis potius rigidis fimbriatae: pagina infera glandibus copiosis subsessilibus flavescentibus; 
pagina superna in juventute glandibus similiter glabrescens; indusium latum tenue 
glandibus obtectum cum vel sine pilis setosis intermixtis. 


Caudex erect. Fronds up to 40 cm long the fertile ones with stipes not conspicuously longer 
than those of the sterile fronds, both stipe and rhachis densely grey-pubescent throughout 
with a mixture of short crisped hairs and some longer, spreading acicular hairs | mm long. 
Lamina 15-30 x 6-8 cm narrowly oblong or elliptic in outline, tapering above, with about 
10-12, less commonly up to 16 pairs of free pinnae, the lowest pair not deflexed. Pinnae 
3-5 x 0'5-1 cm, deeply pinnatifid with up to 10 pairs of oblique, oblong blunt falcate 
segments 2 mm wide, the basal pair quite free and sometimes shortly.stalked, often lobed 
near the base slightly enlarged and lying close to or often underlying the rhachis, sometimes 
the second pair of segments also free to the base, the rest of the pinna becoming progressively 
less deeply cut into oblique, blunt, falcate lobes; apex of pinnae blunt or acute, uppermost 
pinnae and those adnate to the apical portion becoming quite entire. Costas and costules 
with scattered acicular hairs above and below and the margins of the segments fringed with 
rather stiff hairs, under surface with copious, subsessile, pale, yellow glands, upper surface 
with similar glands when young, becoming smooth with.age; indusium Broud, thin, covered 
with glands and with or without intermixed setose hairs, 


Eight miles north-east of Ratnapura, in rocky proiad by stream in aisha ravine, 150 m, 4 Jan. 1951, 
Sledge 808 (BM, holotype; US). Same locality and date, Ballard 1383 (K). 


Shady bank by stream at low elevation near Ratnapura. 
Endemic to Ceylon. 


The new species differs from all other known taxa of Trigonospora in the glandular clothing 
of the pinnae and indusia. The glands have very short stalks and spherical heads. The 
densely felted stipe and rhachis are also distinctive. The lowermost pair of pinna segments is 
quite free as in T. calcarata and T. angustifrons and the second acroscopic segments are also 
sometimes free to the base. 

Fronds which have not been submerged in water are easy to recognise by the copious, 
pale-yellow glands, especially on the under surfaces and indusia. In fronds which have been 
inundated however, easily distinguished by the quantities of grit and other debris lodged in 
the hairy coating of the stipe and rhachis, the glands seem to be absent, suggesting that they 
are easily detached and washed off the leaf surfaces. 


This is the plant referred to as ‘Thelypteris n.sp.’ in Manton & Sledge (Phil. Trans. R. Soc. 
Lond. B, 238: 137 (1954). It was examined cytologically and found to be a diploid with 
n = 36. Only known from the type locality where it was abundant and uniform. 


7. Trigonospora zeylanica (Ching) Sledge, comb. nov. (Fig. 2C) 


Thelypteris zeylanica Ching in Bull. Fan meml Inst. Biol. (Bot.) 6 : 287 (1936). Type: Ceylon, Thwaites 
C.P. 3050 (K). 

Nephrodium (Lastrea) falcilobum var. B Hook., Sp. Fil. 4: 108 (1862). Type: Ceylon, Thwaites C.P. 
3050 (K). 

Lastrea calcarata Bedd., Ferns S. Ind. : 82, t.246 (1864), nom. illeg., non L. calcarata (Blume) Moore 
(1858). Type: Ceylon, Thwaites C.P. 3050 (K). 


THELYPTERIDACEAE OF CEYLON 25 


Aspidium calcaratum var. 8 Thwaites, Enum, Pl. Zeyl.: 391 (1864). Bedd., Handb. Ferns Brit. Ind. : 237 
(1883). 

Lastrea calcarata var. moonii Trimen in Syst. Cat. Fl. Pl. Ferns Ceylon ex Journ. Ceylon Branch Roy. 
Asiat. Soc. : 114 (1885), nom. nud—Dryopteris calcarata var. moonii Trimen in Willis, Rev. Cat. FI. 
Pl. Ferns Ceylon, Peradeniya Manuals 2 : 116 (1911), nom. nud. 


Caudex erect, fronds up to 40 cm long, the fertile ones with stipes not much longer than 
those of the sterile fronds. Lamina 15-30 cm long usually 3-6 cm wide but sometimes up to 
9 cm, narrowly oblong to elliptic oblong in outline with about 15 pairs of free pinnae, the 
lowermost pair not deflexed. Pinnae linear-oblong to oval-oblong or fusiform, 2-5 cm long, 
4-10 mm wide rarely cut half way to the costa into about 8 pairs of oblique falcate lobes, 
often the margins little more than serrate, the distal pinnae entire or subentire; base of 
pinnae narrowly cuneate and often shortly stalked at least in the lower pinnae, with no basal 
pinna segments adjacent to the rhachis; apex of pinnae blunt. Rhachis and upper surface of 
costa hairy and margins of segments sometimes with scattered hairs, otherwise surfaces glabrous 
above and below. Indusia glabrous. 


Collections: Thwaites C.P. 3050 (BM; CGE; K; P; PDA: Hinidoon, Dec. 1853; Singh Rajah Forest, 
April 1855). Thwaites C.P. 992 (BM; CGE; K; P, partim; PDA: Saffragam, Aug. 1821, Moon). Above 
Enselwatte Estate above Deniyaya, shaded places by rivulet, June 1969, Kostermans 23658 (US). 
Kanneliya Forest near Hiniduma, low alt., 7 May 1973, Kostermans 24736 (US). Same locality, 7 Dec. 
1976, Faden 76/504 (K). Sinha Raja Forest near Hedigala, stream side, 75 m, 5 Jan. 1951, Sledge 812 
(BM). Pahale Hewissa, by stream, 30 m, 31 March 1954, Sledge 1370 (BM; K; US). Same locality, 20 
Jan. 1951, Ballard 1522 (K). 


By streams in forest at low elevations in the south of the island. 
Endemic to Ceylon. 


A very distinct species easily recognised by its shallow-lobed or merely serrate pinnae with 
strongly cuneate bases. The pinnae are often shortly stalked but this is not always the case. 
They are, however, always devoid of basal segments lying close to the rhachis; all other 
Ceylonese species of Trigonospora have these basal segments. 

A sheet at Peradeniya, without collector’s name or date, purporting to have come from 
Hakgala, is almost certainly wrongly localised; all other gatherings are from low elevations in 
the southern part of Ceylon. 


X. PSEUDOCYCLOSORUS Ching 


in Acta phytotax. sin. 8 : 322 (1963), emend, Holttum in J/ S. Afr. Bot. 40 : 137 (1974). 


Caudex erect or short-creeping; fronds bipinnatifid with abrupt transition at the base to 
numerous small pinnae and often reduced to swollen tubercles; aerophores present at base of 
lower and reduced pinnae; upper surfaces of costae grooved; veins free, usually raised below, 
the acroscopic basal one passing to the base of the sinus between adjacent pinna segments, 
the basiscopic one reaching the edge above the base of the sinus or the two veins converging 
but not fusing at the sinus base; surfaces glabrous or with acicular hairs, eglandular; sori 
indusiate, sporangia without hairs or glands but usually with a septate hair on the stalk. 
n = 36 (35 also reported). 

About 12 species; three in Africa and nine in tropical and subtropical Asia to Japan and 
Luzon. 


1. Pseudocyclosorus tylodes (Kunze) Ching 


in Acta phytotax. sin. 8 : 323 (1936) [as ‘xylodes’.|—Aspidium tylodes Kunze in Linnaea 24 : 244, 283 
(1851) [‘xylodes’ loc. cit. : 281]; Thwaites, Enum. Pl. Zeyl. : 391 (1864). Type: India, Nilgiris, Schmid 
(B).—Lastrea tylodes (Kunze) Moore, Index Fil. : 107 (1858). Copel., Fern Fl. Philipp. : 330 (1960) 
[as ‘xylodes.’|—Lastrea ochthodes var. tylodes (Kunze) Bedd., Ferns S. Ind. : t.107 (1863); Handb. 


26 W. A. SLEDGE 


Ferns Brit. Ind. : 240 (1883).—Nephrodium prolixum var. tylodes (Kunze) Baker, in Hook.& Baker, 
Syn. Fil. : 268 (1867).—Dryopteris tylodes (Kunze) Christensen, Not. Syst. 1 : 41 (1909) [as ‘xylodes’); 
Index Fil., suppl. 3 : 102 (1934), [as ‘xylodes.’|— Thelypteris tylodes (Kunze) Ching in Bull. Fan meml 
Inst. Biol. (Bot.) 6 : 296 (1935) [as ‘xylodes.”] 


Caudex erect, fronds tufted; stipes up to 50 cm long bearing ovate, brown scales near the 
base, elsewhere glabrous; lamina commonly 40-60 cm long and 10-20 cm wide, sometimes 
100 cm long, oblong-lanceolate in outline, pinnate with 20-30 pairs of sessile narowly 
lanceolate pinnae, in large fronds the pinnae up to 30 x 2:5 cm, in small fronds less than half 
this size, apex of pinnae acuminate, base bearing a dark-coloured aerophore on the 
underside, margins cut down 3 of the way to the costa into falcate, acute segments with 
revolute margins, veins about 10 pairs per segment, prominent on both surfaces, the 
lowermost pair connivent but not confluent at the base of the sinus; lowermost 
pinnae not or little shortened but several pairs of abortive pinnae represented by tubercle-like 
vestiges distributed along the stipe; rhachis hairy above, glabrous beneath; pinnae glabrous 
on both sides save for the costae which are strigose hairy above; texture stiff, subcoriaceous; 
sori inframedial, indusium firm, glabrous persistent. Spores verrucose. 


Collections: Thwaites C.P. 1361 (BM; CGE; K; P; PDA). Newera Eliya, shady woods, Sept. 1844, 
Gardner 1108 (CGE; K). Newera Eliya, 22 Jan. 1908, Bicknell (P). Rambodde, Beckett (BM). Hakgala, 
stream banks in jungle, 28 Feb. 1906, J. C. Willis (PDA). Corbets Gap, by track through jungle, 
1200-1300 m, 9 Dec. 1950, Sledge 551, 571 (BM). Ramboda, bank by roadside, 1575 m, 17 Dec. 1950, 
Sledge 659 (BM). Between Pattipola and Horton Plains, by stream in jungle, 1950 m, 20 Dec. 1950, 
Sledge 669 (BM). Horton Plains, by stream in jungle, 2000 m, 19 Dec. 1950, Sledge 688 (BM). 
Riverstone Estate, Matale District, Central Province, 1100 m, 19 Jan. 1977, Faden 77/183 (K). 
Unlocalised: Freeman 236, 237 (BM). Robinson 160 (K). Walker (K; P). 


In mountain forests of the Central Province above 1000 m. 
Southern India and Ceylon. 


Very large plants may have pinnae 30 cm long but 10-15 cm is a normal size. A small form 
is occasionally met with in which the whole frond scarcely exceeds 30 cm and the pinnae are 
then 4-6 cm long. My specimens (no. 551, 571) from Corbets Gap are such forms and are 
very closely matched by Gamble 12122 and 15317 (K) from the Nilgiris and by two sheets 
from Ceylon at Paris. These small forms bear a superficial resemblance to Trigonospora 
ciliata and have sometimes been misidentified as that species, although the presence of 
abortive pinnae on the stipes and the prominent raised veins are sufficient to distinguish 
them. 

Holttum (in Br. Fern Gaz. 11: 55-56 (1974)) has discussed the alternative spellings 
xylodes and tylodes and shown that the latter epithet (which has a descriptive significance 
lacking in the former) is the spelling which Kunze almost certainly intended. 

Pseudocyclosorus ochthodes (Kunze) Holttum is not a Ceylonese fern. It was recorded from 
Ceylon in Beddome’s Handbook because Thwaites misconstrued Kunze’s Aspidium 
ochthodes, the species referred to under that name in Enum. Pl. Zeyl. : 392 (1864) being 
Amphineuron opulentum (Nephrodium extensum (Blume) Moore). Though this error was 
subsequently corrected and the record withdrawn in Handb. Suppl.:54 (1893), the 
emendation has been overlooked by some later writers. 


XI. AMPHINEURON Holttum 
in Blumea 19 : 45 (1971). 


Caudex erect, decumbent or long-creeping; fronds bipinnatifid the basal pinnae narrowed at 
their bases, not reduced in size; basal veins either free or anastomosing to form a short 
excurrent vein to the base of the sinus between adjacent pinna segments, the rest free; lower 
surface of pinnae bearing short acicular hairs and commonly also subsessile, often yellow 
glands; sori usually confined to lobes of pinnae, usually indusiate; indusia often glandular, 


THELYPTERIDACEAE OF CEYLON oi 


sporangium not bearing hairs or glands near the annulus, stalks commonly bearing a short, 
gland-tipped hair; spores dark, irregularly tuberculate or with irregular thick, branched 
ridges. n = 36. 

About 12-15 species; one widespread in Africa, Mascarene Islands and south-east Asia to 
Australia (northern Queensland) and the Pacific Islands, the rest mainly in south-east Asia, 
Malesia and Melanesia. 


1. Amphineuron opulentum (Kaulf.) Holttum 


in Blumea 19: 45 (1971).—Aspidium opulentum Kaulf., Enum. Fil. Chamisso : 238 (1824). Type: 
Guam, Chamisso (LE). 

Aspidium extensum Blume, Enum. Pl. Jay.: 156 (1828). Type: Java, Pulo Pinang, Blume (L).— 
Nephrodium extensum (Blume) Moore, Index Fil. : 91 (1858). Bedd., Handb. Ferns Brit. India : 269 
(1883).—Dryopteris extensa (Blume) Kuntze, Rev. Gen. Pl. 2: 812 (1891).—Cyclosorus extensus 
(Blume) Ching in Bull. Fan mem. Inst. Biol. (Bot.) 8 : 182 (1938). Holttum, Rev. Fl. Malaya 2 : 264, 
fig. 150 (1955).—Thelypteris extensa (Blume) Morton, Am. Fern J. 49: 113 (1959). Schelpe, Fi. 
Zamb. Pterid. : 193 (1970). 

Nephrodium punctatum Parish ex Bedd., Ferns Brit. India: t.131 (1866). Type: Burma, Moulmein, 
Parish (K). 

Aspidium ochthodes sensu Thwaites, Enum. Pl. Zeyl. : 392 (1864). 
Rhizome creeping; stipes up to 60 cm sometimes longer, clothed near the base with linear 
brown scales, elsewhere minutely hairy; lamina bipinnatifid up to 90 cm long with 15-25 
pairs of pinnae, the lowest pair sometimes much reduced in size; pinnae up to 35 x 3 cm in 
large fronds, half this size in small ones, apex acuminate, base obliquely truncate in the 
upper pinnae becoming cuneate in the lowest pinnae, margins lobed about 3 down to the costa, 
the lobes oblique, slightly falcate with acute apices, margins not ciliate; veins 10-12 
pairs, the /owest veins of adjacent groups sometimes anastomosing below the base of the 
sinus, sometimes meeting at the sinus-base or sometimes passing to the edges of the sinus 
just above the base; upper surfaces of costae strigose-hairy, the costules with a few scattered 
hairs, the veins glabrous or almost so, lower surfaces minutely pubescent on the costae, the 
veins bearing copious small yellow glands; sori supramedial, impressed and forming pustules 
on the upper surface, normally confined to the lobes, the basal 3-5 pairs of veins being sterile 
but occasionally all veins fertile; indusia with small glands on the margins, the surfaces 
glabrous or with a few hairs. 


Collections: Thwaites C.P. 975 (BM; CGE; K; P, in part; PDA). C.P. 990, Trimen in herb. Beddome 
(K). Gardner 1362 (K). Gardner 1106 (sic) (BM; CGE). Heights above Kandy, Nov. 1829, Col. Walker 
(K). Hewelkanduraon Koslanda—Wellawaya road, Monaragalla Dist., Uva Province, c. 400 m, 26 Dec. 
1976. Faden 76/593 (K). Kandy Catchment, in secondary jungle, 750 m., 4 Feb. 1954, Sledge 1094 
(BM). Lady Horton’s Walk, Kandy, in secondary jungle above river c. 600 m, 16 Feb. & 24 March 
1954, Sledge 1142, 1354 (BM). Unlocalised: 1839, Mackenzie (K). Geo. Wall (K; PDA). Ferguson 
(PDA; US). Robinson 161 (K). 
Widely distributed but not very frequent in forests at all altitudes up to 1250 m. 

East Africa, Seychelles, southern India, Ceylon, Burma, Thailand, Malesia, Australia 
(northern Queensland), New Caledonia eastwards to Tahiti and Marquesas. 


The small golden-yellow glands on the veins in both the Ceylonese species of Amphineuron 
are distinctive. In A. opulentum the veins are more plentifully supplied with such glands 
than are those of A. terminans, and the indusia are invariably studded with glands round 
their margins, a feature absent from A. fterminans. The basal pair of veins of adjacent 
pinna-lobes may unite below the sinus or remain free from one another within the same 
pinna. As a rule only the tips of the pinna lobes bear sori in A. terminans, rarely more than 
five of the distal veins on the basiscopic side being fertile, with fewer or none on the 
acroscopic side. In A. opulentum at least seven of the distal veins on both sides of the costule 
are usually fertile, and sometimes all the veins carry sori. 


28 W. A. SLEDGE 


2. Amphineuron terminans (Hook.) Holttum 


in Am. Fern. J. 63:82 (1973).—Nephrodium terminans Hook., Spec. Fil. 4:73 (1862). Bedd., Ferns 
S. India : t.90 (1863). Type: Kumaun, Wallich 386 in Herb. Hook. (K). 

Nephrodium pteroides sensu J.Sm., Cat. Cult. Ferns : 54 (1857). Baker, Syn. Fil. : 289 (1868) p.p. Bedd., 
Handb. Ferns Brit. India :296 (1883), non Polypodium pteroides Retz—Dryopteris pteroides sensu 
Christensen, Ind. Fil. : 287 (1905) p.p., non Polypodium pteroides Retz. 

Dryopteris interrupta sensu Ching in Lingnan Sci. J. 12 : 566 (1933), non Preris interrupta Willd.— 
Cyclosorus interruptus sensu Ching in Bull. Fan memil. Inst. Biol. (Bot.) 8 : 184 (1938) Holttum, 
Rey. Fl. Malaya 2 : 262, fig. 149 (1955), non Pteris interrupta Willd. 


Rhizome wide-creeping; stipes 60 cm or more long, scaly at the base, hairy in the groove, 
becoming glabrescent, rhachis hairy above; lamina up to 90 cm long with up to 25 pairs of 
free pinnae beneath the pinna-like apex, the lowest pair (or two) often reduced to small 
auricles; largest pinnae up to 35 x 2:5 cm but often only half this size, apex long-acuminate, 
base obliquely truncate in upper pinnae becoming attenuate in lower ones, margins lobed 
from | up to 4 way to the costa, the lobes usually broadly rounded to subtruncate above with 
forward-pointing tips or sometimes obliquely triangular and falcate, the margins fringed with 
setose hairs; veins 6-12 pairs, the basal pair anastomosing and forming an excurrent vein to the 
base of the sinus, the next pair passing to the sinus membrane; upper surface of costae strigose- 
hairy, the costules and veins with scattered setose hairs at least in the lobes, lower surface 
minutely pubescent on the costae, costules and veins with small yellow glands on the veins 
especially in the distal parts of the lobes; sori sapramedial on the upper veins always confined to 
the lobes and often to the terminal parts of the lobes; indusia broad, thin, surfaces usually with a 
few hairs, margins eglandular. 


Collections: Thwaites C.P. 990 (BM; CGE; P; PDA). Hantane Range, in forests, July 1844, Gardner 
1106 (CGE; K; P). Common at Kandy, Mrs Chevalier (BM). Heneratgoda, in jungle, 13 July 1927, J. 
M. de Silva (PDA). Rawana ella Falls, Ella-Wellawaya road, Ella Pass, Badulla Dist., Uva Province, c. 
775 m, 18 Nov. 1976, Faden 76/371 (K). Lady Horton’s Walk, Kandy, c. 600 m., 11 Dec. 1950, Sledge 
582 (BM). Kadugannawa, amongst undergrowth below Hevea trees near roadside c. 300 m, 12 Dec. 
1950, Sledge 584 (BM). Hunnasgiriya, c. 870 m, 16 Jan. 1954, Sledge 965 (BM). Ravine south of Bibile, 
Uva Province, 450 m, 22 Feb. 1954, Sledge 1172 (BM). Unlocalised: Robinson C151 (K). Bradford ex 
herb. Hance (BM). Randall in herb. Rawson 3220 (BM). Ferguson (PDA; US). 


Widespread and not uncommon in forests in the west and centre. 
Southern India and Burma to China (Hainan Dao); throughout Malesia to New Guinea 
and Australia (northern Queensland). 


XII. CHRISTELLA H. Lév. 


Fl. Kouy-tchéou : 472 (1915), emend. Holttum in Taxon 20: 533 (1971) and Blumea 19: 43 (1971). 


Caudex erect to wide-creeping; fronds pinnate, decrescent with 1-5 pairs of lower pinnae 
gradually reduced, or rarely the lowest pinnae deflexed but not or scarcely reduced; pinnae 
pinnatifid, lobed or crenate; aerophores not conspicuous; veins anastomosing, rarely free, 
costas, costules, veins and lamina surface bearing erect, acicular hairs and sometimes small 
capitate hairs beneath and sometimes also thick, blunt, orange-coloured glands; sori 
indusiate, sporangia lacking setae or glands near the annulus but always bearing unicellular, 
elongate, glandular hairs on the stalks of the sporangia: spores with incomplete anastomosing 
wings or verrucose. n=36. 

About 50 species, mainly in tropical and subtropical regions of Asia, with smaller 
numbers in Africa and America. 

Holttum’s account of Christella (in Kew Bull. 31 : 293-339 (1976)) is prefaced by a 
discussion of generic characters and of the systematic problems posed by variability and 
hybridisation in the genus. Few groups of ferns were less well understood by Beddome and 
his contemporaries. That variability is, in part, due to hybridisation now seems probable, for 
the readiness with which hybrids may be synthesised experimentally leaves scarcely any 


THELYPTERIDACEAE OF CEYLON 29 


doubt that they must also occur in nature where different species grow intermixed. But the 
extent to which variation is genetically controlled or environmentally influenced is 
unknown, and cytotaxonomic and experimental studies are much needed for further 
progress in the understanding of this difficult genus. Both Manton and A. R. Smith have 
emphasised the occurrence of abortive spores as one of the best ways of recognising hybrid 
specimens. But it does not follow as a corollary that the formation of sound, or apparently 
sound, spores necessarily disproves hybrid origin, and the occurrence of mixed characters in 
some specimens seems irreconcilable with their being of pure stock. 

The spores of Christella are of two types. In one the wall is raised into narrow irregularly 
anastomosing, short and incomplete wings. This type I refer to briefly as ridged. In the other 
type the wall forms more or less flattened, spine-like outgrowths with irregular apices, not 
joined to form interrupted wings. These two types are illustrated by Wood in Bot. J. Linn. 
Soc. 67 (Suppl. 1): 194, pl. 1 A and E (1973). The second type of spore ornamentation, here 
referred to as verrucose, is found in C. papilio and C. meeboldii; other Ceylonese species of 
Christella all have spores of the first type. 

I have felt it unnecessary to repeat all the citations of synonyms given under each species 
in Holttum’s work, especially where these are names originally given to gatherings from 
areas remote from Ceylon. My aim has been to include names given in works dealing with 
Ceylon, India and the western Malayan region. In some instances the references to 
Hooker’ s or Beddome’ s works, as cited in synonymy by Holttum, are differently construed 
by me. 

Seven species are recognised in the following account. Two of the species credited to Ceylon 
by Holttum: Christella malabariensIs and C. taprobanica, | reject for the reasons given 
in the discussions under C. dentata and C. papilio. Only two species, C. parasitica and C. 
dentata are common and widespread. The distribution of C. hispidula and C. papilio is not 
well known; they are infrequent but probably occur in many different stations where suitable 
conditions obtain. Two other species, C. subpubescens and C. meeboldii are at present 
known only from one and two gatherings respectively. Although both may have been 
overlooked through confusion with other species, it is more likely that the absence of 
specimens in all the collections examined gives a true indication of their rarity. The seventh 
species C. zeylanica has not been regathered during the present century. 


Key to species of Christella 


1 One or more pairs of lower pinnae distinctly shortened : , ; : ; 2 
— Lowermost pair of pinnae not or hardly shortened, but often deflexed 1. C. parasitica 
(p. 29) 

2(1) Rhizome erect, frondstufted . . ‘ : : : : : ‘ : : é : 3 

— Rhizomecreeping,frondsspaced  . : : i 4 , ‘ ; : : 4 
3(2) One pair of veins anastomosing : pinnae hairy peas : 3 3. C. hispidula (p. 33) 
- 142 pairs of veins anastomosing : pinnae subglabrous beneath . 6. C. papilio (p. 37) 
4(2) aie subentire, crenate or very shallowly lobed . . 7.C. zeylanica (p. 39) 

—  Pinnae pinnatifid . ; : s : i ’ F : 2 ; : ‘ > 

5(4) Up to 4 pairs of basal piiute SHortened : : F : : ‘ : ; : ; - 6 
— Five or more pairs of basal pinnae shortened . : ‘ 6. C. papilio var. repens (p. 37) 
6(5) At least some pinnae with 2 pairs of veins fusing below the sinus membrane; pinnae lobed 

lessthanhalfwaytocosta .. 1 


- 1-1} pairs of veins fusing below the sinus simmembeanes pinnae lobed page half — to costa 


2. C. dentata (p. 32) 
7(6) Pinnae glabrous or nearly so beneath except on costa . : 4. C. subpubescens (p. 35) 
Smemeionaciharyibeneathi wl So ae ke 5. C. meeboldii (p. 36) 


1. Christella parasitica (L.) H. Lév. (Fig. SA) 


Fl. Kouy-Tchéou: 475 (1915).—Polypodium parasiticum L., Sp. Pl. : 1090 (1753). Type: Canton, 
Osbeck (S—PA, in herb. Swartz).—Nephrodium parasiticum (L.) Desvaux in Mém. Soc. linn. Paris 


30 


W. A. SLEDGE 


\ 
{> ~ 


AN | 
eee Lf 


in I 
7 | ra | \ 
; ero as 


ee ee ee aS Sas = 


| 


) 

i 

\ 
f 
cs 


Fig. 5 Pinna segments: A—Christella parasitica (L.) 
(Decaisne) Holttum (Sledge 578), x 5; C—portion of B enlarged to show sorus and hairs, x 20; 


D-C. dentata (Forssk.) Brownsey & Jermy (Sledge 924), x5; E-C. subpubescens (Blume) 
Holttum (Sledge 585), x 5; F-C. meeboldii (Rosenst.) Holttum (Meebold 2133), x 6. 


6 : 260 (1827). C. B. Clarke in Trans. Linn. Soc. (Bot.) Il. 1 : 533 (1880) p.p.—Dryopteris parasitica 
(L.) Kuntze, Rev. Gen. Pl. 2: 811 (1891). Christensen in Ark. Bot. 9(11): 26, fig. 4 (1910); in Gard. 
Bull. Str. Settl. 4: 389 (1929).—Cyclosorus parasiticus (L.) Farwell in Amer. Midl. Nat. 12 :259 
(1929). Ching in Bull. Fan meml Inst. Biol. (Bot.) 8 : 201 (1938). Holttum, Rey. Fl. Mal. 2: 281, fig. 
162 (1955)—Thelypteris parasitica (L.) Tardieu in Notul. Syst. 7:75 (1938). K. Iwats. in Mem. Coll. 
Sci. Kyoto Uniy. B, 31 : 172 (1965). 

Nephrodium molle Hook., Sp. Fil. 4: 67 (1862) p.p.; Bedd., Ferns S. Ind.: t. 84 (1863); Handb. Ferns Brit. 
Ind. : 277 (1883) p.p. and Handb. Suppl. : 76 (1892) p.p., non A. molle Swartz. —Aspidium molle sensu 
Thwaites, Enum. Pl. Zeyl. : 391 (1864) p.p., non Swartz (1801). 


THELYPTERIDACEAE OF CEYLON 31 


Aspidium procurrens Mettenius in Annis Mus. Bot. Lug.-Bat. 1: 231 (1864). Type: Java, Zippelius (L: 
908, 335-152)—Nephrodium procurrens (Mettenius) Baker, Syn. Fil. : 290 (1867). C. B. Clarke in 
Trans. Linn. Soc. (Bot.) Il. 1: 530 (1880). Bedd., Handb. Ferns Brit. Ind.:278 (1883); Handb. 
Suppl. : 67 (1892) p.p.—Dryopteris procurrens (Mettenius) Kuntze, Rev. Gen. Pl. 2: 813 (1891). van 
Rosenb., Malay. Ferns : 211 (1908) p.p.—Cyclosorus procurrens (Mettenius) Copel., Fern Fl. 
Philipp.: 340 (1960) nomen tantum.—Thelypteris procurrens (Mettenius) C. F. Reed in Phytologia 
17 : 306 (1968). 

Nephrodium didymosorum Parish ex Bedd., Ferns Brit. Ind., t.200 (1866). Type: Burma, Moulmein, 
Parish (not found in Herb. Kew).—Nephrodium molle var. didymosorum (Parish ex Bedd.) Bedd., 
Handb. Ferns Brit. Ind. : 279 (1883).—Dryopteris didymosora (Parish ex Bedd.) Christensen, /ndex 
Fil. : 262 (1905). van Rosenb., Malay. Ferns : 225 (1908) p.p.; idem Suppl. : 183 (1917).—Cyclosorus 
didymosorus (Parish ex Bedd.) Nayar & Kaur, Companion Beddome : 68 (1974). 

Nephrodium tectum Bedd., Handb. Suppl. : 79 (1892), excl. King spec. Type: Singapore, Wallich 394 
(K; W). 


Rhizome creeping; stipes scaly at the base elsewhere more or less hairy according to age, up 
to 40 cm long; rhachis clothed throughout with soft, acicular, unicellular hairs. Lamina 
about as long as stipe or longer, pinnate with up to 20 pairs of pinnae and sometimes more; 
basal pinnae deflexed, not or only slightly reduced. Pinnae 7-15 x 1-2 cm, base truncate, apex 
acuminate, /obed }4 to costa, veins 8-10 pairs in the lobes, the /owermost veins of adjacent 
lobes joining to form an excurrent vein to the sinus, the next pair meeting the edge above the 
base of the sinus or occasionally the second acroscopic vein terminating at the base of the 
sinus; basal acroscopic lobes of lower pinnae somewhat enlarged and normally incurved and 
repand; /ower surface of pinnae covered with sofi spreading, acicular hairs up to 1 mm long, 
and thick orange or yellow glands normally present on the veins in the lobes; upper surface of 
costae densely hairy, fewer hairs on the veins, surface of lamina between the veins with 
scattered short hairs and subsessile colourless, gland-like capitate hairs. Sori medial or 
slightly nearer the margin, rarely more than four pairs of veins in a lobe fertile and often only 
the lowermost veins bearihg sori; indusia bearing long hairs. n = 72 tetraploid. 


Collections: Thwaites C.P. 974 (BM, in part; CGE, in part; K; P, in part; PDA, in part). Peradeniya, 7 
Feb. 1914, Petch (PDA). Roseneath, Kandy, edge of path in secondary jungle, 660 m, 28 Nov. 1950, 
Ballard 1000 (K). Kandy, 720 m, 9 April 1954, Sledge 1412 (BM). Corbets Gap, 1200 m, 7 Jan. 1951, 
Sledge 844 (BM). Nawanagalla, bushy ground by road through jungle, 1110 m, 8 Jan. 1954, Sledge 942 
(BM). Jungle at Hunnasgiriya, 870 m, 16 Jan. 1954, Sledge 977 (BM). Panilkande, 600 m, 5 April 
1954, Sledge 1401 (BM). Galaha, 940 m, 22 Jan. 1954, Schmid 1041 (BM). Badulla road from Nuwara 
Eliya, 29 Dec. 1950, Ballard 1313 (K). Bambaragalla, Matale District, 750 m, 12 Dec. 1976, Faden 
76/530 (K). Unlocalised: Mrs Chevalier (BM). Freeman 270C, 271D, 272E, 273F (BM). Barkly (BM). 
Alston 1093 (K). 
Very common in the Western, Central and Southern Provinces. At all elevations up to 
1800 m, in open or lightly shaded, grassy or bushy places, and roadside banks. 

India, Ceylon, south China and south Japan, through south-eastern Asia to Taiwan, New 
Caledonia and Australia (Queensland) and east to Polynesia. 


Christella parasitica is distinguished from all other Ceylonese species of Christella by its 
non-decrescent fronds, the pinnae of which normally, though not invariably, have orange- 
coloured glands on the lower surface of the veins. Beddome and his contemporaries failed to 
distinguish the species and nearly all the descriptions of taxa in this group in Beddome’s 
Handbook are compounded of two or more species, though his description of Nephrodium 
molle var. didymosorum refers exclusively to Christella parasitica. In the Supplement to his 
Handbook the description of Nephrodium molle includes three species. The first three lines 
refer to Christella parasitica, the next two lines are applicable to C. dentata, and the 
inclusion of C. hispidula is evidenced by the reference to erect rhizomes. His description of 
Nephrodium tectum is also compounded of Christella parasitica and C. hispidula. In the 
Handbook a fourth species is probably included in the description of Nephrodium molle, for 
two pairs of anastomosing veins and a glabrous indusium are the usual characters of 
Christella subpubescens, but not of the other three species which the description covers. 


32 W. A. SLEDGE 


Nephrodium procurrens Baker is construed by Beddome as Christella parasitica, though his 
reference to ‘the very large form . . . with the lower pinnae reduced to deltoid auricles’ clearly 
refers to the fern which Hope later described as Nephrodium papilio. The latter is transferred 
in the Supplement to N. molle var. major, but N. procurrens remains a hybrid description 
covering Christella parasitica and C. cylindrothrix. 


2. Christella dentata (Forssk.) Brownsey & Jermy (Fig. 5D) 


in Brit. Fern. Gaz. 10 : 338 (1973). Holttum, in Jl S. Afr. Bot. 40 : 143 (1974).—Polypodium dentatum 
Forsskal, Fl. Aegypt. Arab. : 185 (1775). Type: Arabia, Forsskal (C).—Dryopteris dentata (Forsskal) 
Christensen in K. dansk. Vidensk. Selsk. Skr. VII, 6(1) : 24 (1920); Index Fil. Suppl. 3 : 84 (1934). 
Backer & Posth., Varenfl. Java : 58 (1939).—Thelypteris dentata (Forsskal) E. St. John in Am. Fern 
J. 26:44 (1936). Schelpe, Fl Zamb. Pterid.: 197 (1970). A. R. Sm., Univ. Calif, Publ. Bot. 59:57 
(1971).—Cyclosorus dentatus (Forsskal) Ching in Bull. Fan mem Inst. Biol. (Bot.) 8 : 206 (1938). 

Polypodium nymphale G. Forster, Fl. Ins. Austr. Prodr.: 81 (1785). Type: New Zealand, G. Forster (BM). 

Aspidium molle Swartz in J. Bot. Gottingen 1800 (2): 34 (1801) [nom. nov. pro Polypodium molle 
Jacq., Collect. Bot. 3: 188 (1789), non Schreber (1771) nec All. (1785)] Type: cult. Vienna (W).— 
Nephrodium molle (Swartz) R.Br., Prodr. Fl. N. Holl. : 149 (1810). Hooker, Sp. Fil. 4 : 67 (1862), p.p. 
Bedd., Handb. Ferns Brit. Ind.:277 (1883) p.p.; Handb. Suppl: 76 (1892) p.p—Dryopteris mollis 
(Swartz) Hieron. in Hedwigia 46 : 348 (1907). van Rosenb., Handb. Suppl: 183 (1917), p.p.? 

Nephrodium malabariense Fée, Mém. Soc. Sci. nat. Strasbourg 6 : 43 (1865). Type: Concan, Law ex 
herb. J. D. Hook. & Thomson (holotype not seen, BM lectotype).—Thelypteris malabariensis (Fée) 
Panigrahi in Notes R. bot. Gdn Edinb. 33 : 497 (1975).—Christella malabariensis (Fée) Holttum, in 
Kew Bull. 31 : 317 (1976) p.p. 

Cyclosorus subpubescens sensu Holttum, Rev. Fl. Mal. 2: 273, fig. 157 (1955), non Aspidium 
subpubescens Blume. Sensu Panigrahi & Manton in J. Linn. Soc. (Bot.) 55 : 729-743 (1958). 

Cyclosorus jaculosus sensu Panigrahi & Manton /oc. cit., non Aspidium jaculosum Christ. 

Thelypteris taprobanica Panigrahi in Kew Bull. 31 : 187 (1976) p.p., incl. holotype. Type: Ceylon, G. 
Wall (K). 


Rhizome shortly creeping, fronds dimorphic, the fertile ones with more remote pinnae and 
standing high above the more spreading sterile fronds; stipes variable in length, scaly at the 
base elsewhere hairy as is the rhachis. Lamina longer than the stipe, pinnate with 15-25 pairs 
of free pinnae below the pinnatifid apex, Jower 1-4 pairs reduced and widely spaced, the 
basal acroscopic lobes of the lower pinnae enlarged, often with lobed margins, its veins 
forked and sometimes anastomosing. Pinnae up to 14x 1:5cm, base truncate, apex 
acuminate, lobed about half way or a little more to the costa, veins 7-9 pairs in the lobes, the 
lowermost veins of adjacent lobes joining to form an excurrent vein to the sinus, the second 
acroscopic vein and sometimes the basiscopic vein also, reaching the excurrent vein at or 
about its confluence with the sinus membrane; Jower surface of pinnae eglandular, normally 
pubescent with short hairs on the veins and very short hairs on the lamina surface, but 
sometimes almost glabrous, a few longer acicular hairs on the costa mixed with the 
prevailing shorter ones; upper surface of costae densely hairy, veins with short stiff hairs and 
some longer acicular hairs, lamina surface with or without short hairs, subsessile capitate 
hairs absent. Sori medial, four to six pairs of veins commonly fertile; indusia normally 
short-hairy, but sometimes glabrous. n = 72 tetraploid. 


Collections: Thwaites C.P. 714 (BM; CGE; P). Thwaites C.P. 974 in part (BM; CGE; P; PDA). 
Thwaites C.P. 3498 (PDA, one of four sheets). Hantane Range, in forest, July 1844, Gardner 1105 
(CGE). Gardner 1105 ex herb. J. Smith (BM). Lady Horton’s Walk, Kandy, Robinson 148 (K). Hakgala, 
1800 m, 23 Dec. 1950. Sledge 704 (BM). Hakgala Peak, sub C. jaculosus, Dec. 1950, Manton P199 (BM). 
Same locality, 1700-1800 m, 14 Nov. 1976, Faden 76/271 (K). Nuwara Eliya, Freeman 267A, 268A, 
269B (BM). Jungle at Henaratgoda, 14 July 1927, J. M. de Silva (PDA). Kandy, in jungle, 3 June 1927, 
Alston 1093 (K; PDA). Corbets Gap, 1200 m, 7 Jan. 1951, Sledge 834 (BM). Two miles east of Panilkanda, 
edge of jungle near road, 660 m, 24 Jan. 1951, Sledge 924 (BM). Same locality & date. Ballard 1563 (K). 
Same locality, 5 April 1954, Sledge 1402 (BM). Hunnasgiriya, jungle 870 m, 16 Jan. 1954, Sledge 978 (BM). 
Le Vallon, forest, c. 1500 m, 9 Feb. 1954, Sledge 1128 (BM). Panadura, roadside in coconut plantation, 
2-4 m, 14 Feb. 1968, Comanor 997 (K; PDA). Kadugannawa, on damp bank in rubber plantation, 12 Dec. 


4 


THELYPTERIDACEAE OF CEYLON 33 


1950, Ballard 1091 (K). Bambragalla, Matale District, 750 m, 12 Dec., 1976, Faden 76/531 (K). Near 
Urugala, on Kandy-Mahiyangana road, 650 m, 24 Dec. 1976, Faden 76/562 (K). Bakinigahawela on Bibile- 
Moneragala road, shady stream bank, c. 250 m, 25 Dec. 1976, Faden 76/573 (K). 


In similar situations to and as widespread as Christella parasitica. 
Tropics and subtropics of Africa, Asia and Polynesia south to New Zealand. Introduced in 
America. 


In the absence of rhizomes, the fronds of Christella dentata may be distinguished from those 
of C. hispidula by their venation. The second pair of veins in C. hispidula invariably reaches 
the margins of the lobes well above the base of the sinus (which lacks a distinct membrane), 
whereas in C. dentata one or both run into the sinus membrane and normally the second 
acroscopic vein fuses with the excurrent vein at or about the base of the sinus. The fronds 
also lack the small, colourless, gland-like, capitate hairs which occur on the upper surface of 
the pinnae in C. hispidula. They differ from C. parasitica in being dimorphic, decrescent and 
eglandular, the pinnae being short-hairy to subglabrous beneath and never with abundant 
long, soft, acicular hairs. C. subpubescens differs in its shallowly lobed, subglabrous pinnae 
in which two pairs of veins anastomose below the sinus. 

Fée’s description of Nephrodium malabariense includes no differences from Christella 
dentata other than minor quantitative distinctions, Holttum enlarged Fée’s description to 
cover glandular plants, and both authors wrongly attributed the type locality to Malabar (see 
Sledge in Kew Bull. 34 : 78 (1979)). 

The holotype of Panigrahi’s Thelypteris taprobanica is a large, almost glabrous, form of 
Christella dentata. 


3. Christella hispidula (Decaisne) Holttum (Fig. 5B, C) 


in Kew Bull. 31: 312 (1976).—Aspidium hispidulum Decaisne in Nouy. Ann. Mus. Hist. nat. Paris 
3 : 346 (1834). Type: Timor, Guichenot (P).—Nephrodium hispidulum (Decaisne) Baker in Hook. 
& Baker, Syn. Fil. : 293 (1867) nomen tantum—Dryopteris hispidula (Decaisne) Kuntze, Rev. 
Gen. Pl. 2: (1891). Christensen, Jndex Fil. : 271 (1905), excl. syn. Nephrodium angustifolium C. 
Presl, & N. smithianum C. Presl. van Rosenb., Malay. Ferns : 228 (1909), excl. syn. N. angustifolium 
C. Presl.—Cyclosorus hispidulus (Decaisne) Ching in Bull. Fan meml Inst. Biol. (Bot.) 10: 245 
(1941).—Thelypteris hispidula (Decaisne) C. F. Reed in Phytologia 17 : 283 (1968). 

Nephrodium hilsenbergii C. Presl, Epim. Bot. : 47 (1851). Type: Mauritius, Sieber Syn. Fil. 49 (PRC, 
holotype; K, P, isotypes).—Christella hilsenbergii (C. Pres!) Holttum in J/ S. Afr. Bot. 40: 142 
(1974).—Thelypteris hilsenbergii (C. Presl) Panigrahi in Phytologia 31 : 369 (1975). 

Nephrodium quadrangulare Fée, Gen. Fil. : 308 (1852). Type: Guyana, Leprieur 182 (?P, holotype; 
NY, isotype).— Dryopteris quadrangularis (Fée) Alston in J. Bot. Lond. 75 : 253 (1937).—Cyclosorus 
quadrangularis (Fée) Tardieu, Phanérogamie 14: 345 (1952)—Thelypteris quadrangularis (Fée) 
Schelpe in J/ S. Afr. Bot. 30: 196 (1964); 31 : 264, t.1 fig. b (1965); Fl. Zamb. Pterid. : 195 (1970). 
A.R.Sm., Univ. Calif. Publs. Bot. 59 : 64, fig. 114 (1971). 

Nephrodium tectum Bedd., Handb. Suppl. : 79 (1892) p.p. quoad King spec. ex Perak. 

Nephrodium molle sensu Bedd., Handb. Ferns Brit. Ind. : 277 (1883) p.p., Handb. Suppl. : 76 (1892) 
p.p., non Aspidium molle Swartz. 

Dryopteris contigua Rosenst. in Meded. Rijks. 31 : 8 (1917). Christensen in Gard. Bull. Str. Settl. 7 : 244 
(1934). Type: Borneo, Teuscher (L.).—Cyclosorus contiguus (Rosenst.) Ching in Bull. Fan meml 
Inst. Biol. (Bot.) 10: 243 (1941). Holttum, Rev. Fl. Mal. 2: 282, fig. 163 (1955).—Thelypteris 
contigua (Rosenst.) Reed in Phytologia 17 : 269 (1968). 

Dryopteris repandula van Rosenb. in Nova Guinea 14:20 (1924). Type: western New Guinea, 
Mamberamo River, 90m, Lam 1058.(BO, holotype; L, isotype}—Cyclosorus repandulus (van 
Rosenb.) Ching in Bull. Fan meml Inst. Biol. (Bot.) 10 : 248 (1941). Panigrahi & Manton in J. Linn. 
Soc. (Bot.) 55 : 729-743 (1958).—Thelypteris repandula (van Rosenb.) Reed in Phytologia 17 : 308 
(1968). 


Rhizome erect, fronds tufted; stipes usually short, 5 cm long, but sometimes 20 cm or more, 
stramineous, scaly at the base, elsewhere beset with pale, slender hairs; rhachis similarly 


34 W. A. SLEDGE 


clothed throughout with acicular hairs and densely so when young. Lamina (20) 30-60 
(75) cm pale green or yellow-green in colour and softly herbaceous in texture, normally 
pinnate throughout with no terminal pinna-like part and with up to 30 pairs of pinnae; 
decrescent below with up to four pairs of pinnae shortened and deflexed and the basal 
acroscopic lobes of the middle and lower pinnae enlarged, incurved and more or less repand. 
Pinnae 7-15 x 1-2 cm, base truncate, apex acuminate, margins lobed 3 to costa, veins 7-9 
pairs in the lobes, the lowermost veins of adjacent lobes joining to form an excurrent vein to 
the sinus (which lacks a distinct basal membrane), the next pair meeting the edge above the 
base of the sinus; lower surface of pinnae with a mixture of acicular hairs 1-1°5 mm long and 
shorter hairs on the costae and mainly short hairs on the veins and lamina surface in the 
lobes; upper surface of costae with many acicular hairs, similar hairs occurring more sparsely 
on the veins and a mixture of short hairs and minute, colourless, subsessile, gland-like 
capitate hairs on the lamina surface. Sori medial or supramedial, rarely more than five pairs 
of veins in a lobe fertile and sori often present only on the basal veins; indusia normally 
sparsely pilose with long hairs but sometimes almost or quite glabrous. n = 36, diploid. 


Collections: Thwaites C.P. 974 in part (CGE; PDA). Thwaites C.P. 714 (K; PDA, in part). Heights 
above Kandy, Walker in Herb. Hook. (K). Lady Horton’s Walk, Kandy, 600.m, 11 Dec. 1950, Sledge 
578 (BM). Same locality, 6 Feb. 1954, Sledge 1100, 24 March 1954, Sledge 1356 (BM). Roseneath 
Valley, Kandy, 630 m, 3 Feb. 1954, T. G. Walker, Sledge 1092 (BM). Kandy, Catchment on bank in 
secondary jungle, 750 m, 4 Feb. 1954, Sledge 1098 (BM). Laxapana, 900 m, 28 Jan. 1954, Sledge 1060 
(BM). Allagalla, 600 m, 19 Feb. 1954, Sledge 1151 (BM). Hillside above Potupitya, Ratnapura District, 
forest patch, much degraded, 1500 m, 4 Dec. 1976, Faden 76/480 (K). Between Kirapatdeniya & 
Weligepola, Ratnapura District, swampy places in forest patch above road, 1900-2000 m, 30 Dec. 
1976, Faden 76/653 (K). Unlocalised: 1899, Bradford (K). 1861, Hance 132 ex Thwaites (P). Ferguson 
[as Nephrodium molle] (US 816419). 


Distribution imperfectly known but probably widespread; less frequent than Christella 
dentata and C. parasitica. 
Tropics of Asia, Africa and America. 


Holttum’s description of Christella hispidula states ‘Rhizome erect or short-creeping.’ I 
believe the rhizome is a/ways erect and that the apparently creeping habit displayed by some 
herbarium specimens 1s illusory. Two of my gatherings (Sledge 1092, 1098) would certainly 
be construed by most botanists as having creeping rhizomes; yet these were taken from a very 
steep clay bank by a road through forest. In such a position soil movement tends to cover and 
depress the rhizome into a horizontal position, the apex on emergence turning upwards 
before unfurling the tufted fronds, the excavated rhizomes thus having every appearance of 
being horizontally creeping. C. hispidula commonly grows in clearings or where 
circumstances allow increased light penetration, as where roads run through jungle and 
forest; thus there must be other gatherings which, like my own, display a similar stem apex 
and fronds. 

Christella hispidula is easily recognised by its habit of growth and straw-coloured, hairy 
stipes and decrescent fronds. Living fronds are normally pale green or yellow-green, but 
there is little evidence in dried fronds of this colour difference. The fronds vary in size, often 
the stipes being short with crowded lower pinnae, but they may be over 20 cm long with the 
lower pinnae well spaced. They are also pinnate to the apex, rarely terminating in a 
pinnatifid pinna-like part as in most other Ceylon species of Christella. The didymosoral 
condition produced by only the basal pair of veins being fertile and hence giving rise to a 
single row of paired and closely contiguous sori on each side of the costa, commonly occurs in 
both C. hispidula and C. parasitica, but I have not observed it in other Ceylon species of 
Christella. 

Beddome failed to distinguish C. hispidula, C. parasitica and C. dentata (see p. 31) and 
therefore attributed to Nephrodium molle a rhizome which may be either erect or creeping 
and fronds which may or may not be decrescent. 


THELYPTERIDACEAE OF CEYLON 35 
4. Christella subpubescens (Blume) Holttum (Fig. 5E) 


in Webbia 30 : 193 (1976).—Aspidium subpubescens Blume, Enum. Pl. Jay. : 149 (1828). Type: Java, 
Blume (L:910, 327-113, large sheet collection)—Dryopteris subpubescens (Blume) Christensen in 
Gard. Bull. Str. Settl. 4: 390 (1929); Index Fil. Suppl. 3 : 99 (1934). Backer & Posth., Varenfl. Java: 
65 (1939) p.p— Cyclosorus subpubescens (Blume) Ching in Bull. Fan memil Inst. Biol. (Bot.) 8: 211 
(1938) p.p.—Thelypteris subpubescens (Blume) K. Iwats. in Mem. Coll. Sci. Kyoto Univ. B, 31: 173 
(1965), excl. syn. Aspidium jaculosum Christ. 

Aspidium amboinense sensu Blume, Enum. Pl. Jav.: 148 (1828), non Willd. Kunze in Bot. Ztg. 6 : 261 
(1848). Mettenius, Farngatt. pt IV : 105 (1858).—Nephrodium amboinense sensu Hook., Spec. Fil. 
4:75 (1862); Syn Fil. : 292 (1867) p.p.; Bedd., Suppl. Ferns S. Ind. & Brit. Ind. : 19 (1876); Handb.; 
Suppl. : 75 (1892) p.p—Nephrodium molle var. amboinense sensu Bedd., Handb. Ferns Brit. Ind. : 278 
(1883), excl. syn. Nephrodium extensum var. minor Bedd. 

Aspidium molle var. latipinna Benth., Fl. Hongkong. : 455 (1861). Lectotype: Hong Kong, Hance 135 
(K).—Nephrodium latipinna (Benth.) Hook., in Hook. & Baker, Syn. Fil. : 292 (1867).—Aspidium 
latipinna (Benth.) Hance in J. Linn. Soc. (Bot.) 13: 141 (1873).—Dryopteris latipinna (Benth.) 
Kuntze, Rev. Gen. Pl. 2: 813 (1891). van Rosenb., Handb. Malay. Ferns : 217 (1908).—Christella 
latipinna (Benth.) H. Lév., Fl. Kouy-tchéou : 475 (1915).—Cyclosorus latipinna (Benth.) Tardieu in 
Phanérogamie 7 : 73 (1938). Tardieu & Christensen in Lecomte, F/. Gén. Indoch. 7 (2): 397 (1941). 
Holttum, Rev. Fl. Mal. 2 : 276, fig. 159 (1955).—Thelypteris latipinna (Bentham) K. Iwats. in Acta 
phytotax. geobot. Kyoto 21 : 166 (1965). Morton in Contr. U.S. natn. Herb. 38 : 361 (1974). 

Dryopteris sumatrana van Rosenb., Malay. Ferns : 227 (1908) p. maj. p. Type: Sumatra, 1778, C. 
Miller (BM).—Cyclosorus sumatranus (van Rosenb.) Ching in Bull. Fan meml Inst. Biol. (Bot.) 
10: 249 (1941). Holttum, Rev. Fl. Mal. 2: 275, fig. 158 (1955).—Thelypteris sumatrana (van 
Rosenb.) Tagawa & K. Iwats. in Acta Phytotax. Geobot. Kyoto 22: 101 (1967). 

Nephrodium molle var. major Bedd., Handb. Suppl.:76 (1892) quoad pl. Sumatr. tantum.— 
Dryopteris subpubescens var. major (Bedd.) Christensen in Gard. Bull. Str. Settl. 4:390 
(1929). 

Dryopteris pseudoamboinensis Rosenst., Meded. Rijks. 31 : 7 (1917). Lectotype (selected by Panigrahi): 
Sumatra, Korthals 270 (L.).—Thelypteris pseudoamboinensis (Rosenst.) Panigrahi in Phytologia 
30 : 410, pl. 111 (1975). 

Thelypteris blumei Panigrahi in Phytologia 30 : 409 (1975); op. cit. 31 : 369 (1975). 


Rhizome short-creeping; stipes up to 30cm long, scaly below, thinly pubescent above; 
rhachis hairy on the upper surface, subglabrous beneath. Lamina up to 80 cm long, pinnate 
with up to 20 pairs of free pinnae below the pinnatifid apex, lower 24 pairs gradually 
reduced and widely spaced, the basal acroscopic lobes of the lower pinnae enlarged with 
forked and often anastomosing veins. Pinnae up to 12x 14cm, base truncate, apex 
acuminate, lobed from | to less than half way to costa, veins 4-8 pairs in the lobes, excurrent 
vein from the union of the lowermost veins of adjacent lobes joined below or at the base of the 
sinus by the second pair of veins, the next acroscopic vein often reaching the base of the sinus 
in large pinnae; /ower surface of pinnae with scattered short hairs mainly on the costae and 
veins; upper surface of costae and veins with short acicular hairs, elsewhere surfaces 
glabrous. Sori medial; indusia normally glabrous but often short-hairy. n = 72 tetraploid. 


Collection: Kadugannawa, amongst undergrowth below Hevea trees near roadside, 300 m, 12 Dec. 
1950, Sledge 585 (BM). 


Apparently very rare but perhaps overlooked through confusion with other species. 
Also in north-east India to south China, Burma, Thailand, Vietnam, Malaysia to New 
Guinea, Philippines and Fiji. 


Christella subpubescens differs from C. dentata and allied species in its less deeply divided 
and less hairy pinnae. I have not been able to detect the very short capitate hairs which 
Holttum says are usually present on the lower surfaces of the pinnae and the very minute 
hairs—best seen with a binocular—which seem normally to be present on the lamina surface in 
this species are not evident in my specimen. The indusia of my gathering are also glabrous, 
but they are hairy in an example sent from Singapore by Holttum. The pinnae are more 


36 W. A. SLEDGE 


deeply lobed than in most examples of C. subpubescens, but in their venation and other 
respects they are typical. Holttum has seen the gathering and confirmed the identification. 

Beddome cited no Ceylon stations for Nephrodium amboinense, which is not figured in 
his two volumes of illustrations, but it should be noted that in the Supplement to the Ferns of 
Southern India and British India. : 19 (1876) he distinguished between C. subpubescens and 
C. zeylanica—as Nephrodium amboinense and N. amboinense var. minor respectively— 
giving the distribution of the former as north India and Ceylon and of the latter as Ceylon 
only. Later, in his Handbook and its Supplement, the two were no longer treated as distinct 
from one another. 

In his Ferns of Malaya (Rev. Fl. Mal. 2: 276 (1955)) Holttum states that Cyclosorus 
latipinna ‘certainly occurs in Assam and Ceylon’ though in his paper on Christella (Kew 
Bull. 31: 293-339 (1976)), where C. Jatipinna is treated as synonymous with C. 
subpubescens, Ceylon is not included in the distribution. It is probable that the species 
occurs elsewhere than at Kadugannawa although it must be rare since I have seen no other 
gathering amongst the extensive herbarium material examined. It has not been recorded 
from southern India and the distribution attributed to it by Holttum suggests that in Ceylon it is 
at the westernmost extremity of its range. 


5. Christella meeboldii (Rosenst.) Holttum (Fig. 5F) 


in Nayar & Kaur, Companion Beddome : 208 (1974); Fl. Hassan Dist. : 860 (1976).—Dryopteris 
meeboldii Rosenst. in Reprium Spec. nov. Regni veg. 12 : 247 (1913). Christensen, /ndex Fil. Suppl. 
2:15 (1917). Type: Southern India, Tellicherry, Meebold 2133 (WRSL).—Cyclosorus meeboldii 
(Rosenst.) Ching in Bull. Fan meml Inst. Biol. (Bot.) 8 : 210 (1938), excl. Stocks & Law spec. ex 
‘Malabar’.—Thelypteris meeboldii (Rosent.) C. F. Reed in Phytologia 17: 291 (1968). 

Christella malabariensis (Fée) Holttum in Kew Bull. 31 : 317 (1976) p.p. 


Rhizome short-creeping, clothed about the apex and on the lower parts of the stipes with 
lanceolate, long-attenuate, brown scales; fronds 30-60 cm long, up to 15 cm broad, pinnate 
with about 15 pairs of free sessile pinnae below the narrowly deltoid, deeply lobed apex, 
lower 2-4 pairs reduced in size and deflexed, basal acroscopic lobes of lower pinnae slightly 
enlarged; largest pinnae 5-8°5 x 1-1-5 cm, lobed from } to nearly half way to the costa, the 
lobes broadly rounded to truncate with forward-pointing tips, veins 5-8 pairs in the lobes 
14-2 pairs anastomosing, apex of pinnae entire, acute or acuminate, base truncate or 
obliquely truncate; upper part of stipe and rhachis with spreading hairs, ower surface of 
pinnae with short, spreading hairs on the costa, costules, veins and margins of segments and 
a few longer, acicular hairs on the costa, upper surface hairy on the costa and veins, minutely 
hairy on the surfaces and, at least when young, with scattered, short, capitate hairs especially 
in the basal regions of the pinnae; sori medial to supramedial, the lowest on veins from 
adjacent costules often touching or merging; indusia glabrous or with a few hairs; spores 
verrucose. 

Collections: Near Badulla road [from Hakgala], open ground, 29 Dec. 1950, Ballard 1315 (K). Near 
Parawella Falls, Kandapola, Nuwara Eliya, 1450 m, 19 March 1954, Sledge 1327 (BM). 


Distribution in Ceylon not known; probably of rare occurrence. 
Also in southern India. 


In his account of the genus Christella, Holttum reduced C. meeboldii to a synonym of C. 
malabariensis. | have shown elsewhere (Kew Bull. 34: 77-81 (1979)) that Fée’s Nephrodium 
malabariense is the same as Christella dentata, though Holttum included dentata-parasitica 
hybrids in his concept of C. malabariensis. However, Meebold’s gathering from southern 
India, on which Rosenstock based his species, had not been seen by Holttum for his paper. 
An isotype was subsequently traced to and obtained from Wroclaw, and a photograph and 
description of this were sent to me. Ching had also treated Meebold’s gathering, and the 
Stocks and Law specimens on one of which Fée based his Nephrodium malabariense, as 
belonging to the same taxon, for which however he used the name Cyclosorus meeboldii. I 


THELYPTERIDACEAE OF CEYLON a 


followed Holttum and Ching in my paper on the identity of Fée’s species in not treating 
Christella meeboldii as distinct from the Stocks and Law plants. 

My Kandapola gathering, which differs from all other Ceylon species save Christella 
papilio in having verrucose spores, had been set aside as a presumed new species. The 
shallow lobing of the pinnae, however, often with two pairs of veins anastomosing well 
below the sinus membrane, matched that of Meebold’s south Indian plant. Since no 
information was available about the spores of C. meeboldii, the Wroclaw specimen was again 
sent for and this also proved to have verrucose spores. This both fixed the identity of my own 
plant, whilst providing a sure way of distinguishing C. meeboldii from all other species of 
Christella in Ceylon, for although it has the same type of spore ornamentation as in C. 
papilio, there is no similarity between the fronds of these species. 

Forms of Christella dentata occur in which the pinnae are cut less than half way down and 
such forms approach C. meeboldii in appearance, though they rarely if ever have two pairs of 
veins fully confluent below the sinus membrane. In C. meeboldii the uppermost sori in the 
pinna segments are medial in position, and successive pairs tend to diverge downwards so 
that the sori on the lowermost acroscopic and basiscopic veins of adjacent segments are 
closely contiguous, and hence the rows of sori in fully fertile fronds form an inverted 
V-shape. Rarely there may be a single sorus at the point of anastomosis or it may even be 
inserted onthe excurrent vein. The indusia are sparsely hairy or glabrous. Though some of 
these characters may be found in the variable C. dentata, they are never combined in any 
form of that species. 

From Christella subpubescens, which it most closely resembles in its shallow lobing and 
venation C. meeboldii differs in its hairy fronds. Fortunately the distinctive spores of C. 
meeboldii afford a means of identification if macroscopic characters remain doubtful. 


6. Christella papilio (Hope) Holttum 


in Nayar & Kaur, Companion Beddome : 208 (1974).—Nephrodium papilio Hope in J. Bombay nat. 
Hist. Soc. 12 : 625, t.12 (1899). Lectotype: India, Darjeeling, 1880, Levinge (K).—Dryopteris papilio 
(Hope) Christensen, /ndex Fil. : 282 (1905).—Cyclosorus papilio (Hope) Ching in Bull. Fan meml 
Inst. Biol. (Bot.) 8 : 214 (1938).—Thelypteris papilio (Hope) K. Iwats. in Mem. Coll. Sci. Kyoto Univ. 
B, 31 : 175 (1965). 

Aspidium extensum sensu Thwaites, Enum. Pl. Zeyl. : 391 (1864) quoad C.P. 3498, non Blume. 

Nephrodium molle var. major Bedd., Handb. Suppl. : 76 (1892)excl. pl. ex Sumatra. 


Rhizome erect or (var. repens) short-creeping; stipes up to 20 cm, stramineous, scaly at the 
base, glabrous above. Fronds (including stipes) up to 120 cm, decrescent, pinnate with about 
20 pairs of free pinnae beneath the pinnatifid terminal pinna, largest pinnae commonly 
10 x 1-5 cm but up to twice as long and proportionately wider, caudate-acuminate, lobed 
from one third to half way to the costa, lower 5-12 pairs of pinnae much shortened, well 
spaced, mostly broadly triangular in outline with basal acroscopic lobes enlarged and 
sometimes free almost to the base; veins 6-9 pairs in the lobes 142 pairs anastomising, 
texture herbaceous; surfaces apparently glabrous save on the costa above but minute hairs 
present on the lamina surface above and below, a few scattered short hairs on the veins and, 
very rarely, sessile yellow glands beneath. Sori medial; indusia usually short-hairy, 
sometimes glabrous; spores verrucose. 

var. papilio Rhizome erect. Diploid. 


Collection: Thwaites C.P. 3498 (PDA). Sub Nephrodium amboinense, on same sheet as C. zeylanica 
(PDA). 


var. repens Sledge, var. nov. 


Thelypteris taprobanica Panigrahi, Kew Bull. 31 : 187 (1976) p.p. Type: Ceylon, Wall (K, isotype). 
Rhizoma repens : chromosomatum numerus tetraploideus. 


38 W. A. SLEDGE 


Collections: Hakgala, by jungle stream, 1650 m, 27 Dec. 1950, Sledge 744 (BM). Below Hakgala 
Gardens, jungle, 1650 m, 26 Feb. 1954, Sledge 1212 (BM, holotype). Tangamalai Sanctuary, Haputale, 
1500 m, 25 Feb. 1954, Sledge 1205 (BM) Unlocalised: Wall (K). 

Specimens lacking rhizomes: Thwaites C.P. 3498 (BM; CGE; K; P; PDA). 1819, Moon (BM). 
Ambawella, Wall (K), Ferguson (US 826085). 


By streams in forest in the higher parts of Central and Uva Provinces. Also in southern India 
and north-west Himalayas eastwards to Sikkim, Thailand, northern Malaysia and Taiwan. 


Hope’s species, as described from north-west India, is readily recognised by its erect, and 
often emergent, rhizomes and its almost glabrous, strongly decrescent fronds in which the 
lower pinnae are reduced progressively to short, deltoid, auricled appendages arranged in 
opposite, and hence butterfly-like, pairs. Hope included Ceylon in the distribution of 
Christella papilio, accepting specimens of Thwaites C.P. 3498 and of G. Wall as identical 
with his new species, though these gatherings are lacking rhizomes. Beddome (in Handb. 
Suppl. : 76 (1892) had stated that ‘Mr Wall’s Ceylon specimens however have a decidedly 
creeping root’ and when Hope saw the Ceylon specimens to which Beddome referred, he 
expressed the view that despite the similarity of their fronds ‘the creeping rhizome I consider 
quite enough to separate them... from N. papilio.’ 

Very few herbarium specimens from Ceylon include rhizomes and where these are 
present they are short-creeping, as in my own three gatherings. There are, however, in the 
Peradeniya collection, two specimens which plainly show vertically erect rhizomes, and 
plants with erect rhizomes also occur in southern India (Gamble 11698 and 11769, Nilgiris 
District, 1883 (K)). The Peradeniya specimens are the only ones from Ceylon with erect 
rhizomes seen by me in the numerous sheets examined from many herbaria. 

Manton (in Manton & Sledge, Phil. Trans. R. Soc. 238 : 138 (1954)) found that a Ceylon 
plant (with creeping rhizome) was tetraploid (n= 72), whereas Loyal (in Proc. 48th Indian 
Sci. Congr. Pt. 3 : 266-267 (1961)) has shown that plants from east and west Himalayas 
(presumably with erect rhizomes) are diploid (n = 36). I can find no frond or spore characters 
consistently linked with an erect or with a creeping rhizome which would serve as a means of 
distinguishing between diploid and tetraploid in the absence of basal parts. The inclusion of 
Ceylon moreover within the distributional range of Christella papilio by Ching (in Bull. 
Fan mem Inst. Biol. (Bot.) 8 : 24 (1938)) and Holttum (Kew Bull. 3 : 322 (1976)) (underlines 
the fact, first recognised by Hope, that it is not possible to distinguish isolated fronds derived 
from plants with the different growth habits. 

In the same issue of the Journal of the Bombay Natural History Society as that in which 
Hope described Nephrodium papilio, another new species Polypodium late-repens is 
described, differing from the erect-growing P. distans D. Don (= Pseudophegopteris 
pyrrhorhachis (Kunze) Ching) in its widely creeping and branched rhizome. But Holttum 
(Blumea 17 : 24 (1969)) treats the former as a synonym of the latter on the grounds that ‘I can 
see no distinction between specimens lacking rhizomes.’ The situation is closely parallel in 
the case of Christella papilio from northern India and Ceylon, save for the known 
cytological difference, and this in conjunction with the difference in growth habit justifies 
nomenclatural recognition at varietal level, although as most herbarium specimens lack 
basal parts, only the aggregate name can be given. 

The position has recently been confused rather than clarified by Panigrahi’s (Kew Bull. 
31 : 187-188 (1976)) description of Thelypteris taprobanica. Knowing that Ceylon plants 
differ from Hope’s Christella papilio in their creeping habit and in their cytology, he 
evidently extracted from the Ceylon C. papilio cover at Kew the three specimens which 
showed a creeping rhizome and based his new species on these specimens. However, two of 
these, including his holotype, should not have been included under C. papilio for they are, in 
fact, specimens of C. dentata. They both lack the numerous and characteristic triangular- 
deltoid, reduced, auriculate and butterfly-like lower pinnae and both have the ridged spore 
ornamentation of C. dentata. The third specimen (Panigrahi’s ‘isotype’) is manifestly 
different, having the short, broad lower pinnae of C. papilio and verrucose spores as 


THELYPTERIDACEAE OF CEYLON 39 


illustrated by Wood in Bot. J. Linn. Soc. 67 (Suppl. 1): 194, pl. LE (1973). Panigrahi’s 
description moreover ’whole frond glabrous’ and ‘indusium glabrous’ is applicable to only 
one of the three specimens; two have hirsute indusia and one is hairy on the lower surface. 
Holttum overlooked Panigrahi’s error and has recognised his species as Christella 
taprobanica in his account of the genus. 

Holttum (loc. cit.) states that specimens of Christella papilio from Thailand and Malaya 
may be glandular on the lower surface of the fronds, but no such glandular specimens had 
been seen by him from India. Ferguson’s specimen in US., named Nephrodium amboinense 
and misidentified by Ching and Iwatsuki as Cyclosorus subpubescens, is studded below with 
abundant, sessile, yellow glands. I have seen no other glandular specimen of C. papilio from 
Ceylon. 


7. Christella zeylanica (Fée) Holttum 


in Nayar and Kaur, Companion Beddome : 208 (1974).—Nephrodium zeylanicum Fée, Mem. Fam. 
Foug.10 : 42 (1865). Type: Ceylon, Thwaites C.P. 3391 (holotype not seen; BM, CGE, K (errore 
3390), PDA, isotypes). 

Aspidium extensum sensu Thwaites, Enum. Pl. Zeyl. : 391 (1864) quoad C.P. 3391 non Blume. 

Nephrodium extensum var. minor Bedd., Ferns Brit. Ind. : t.201 (poor) (1866). Type: Ceylon, Thwaites 
(K).—Nephrodium amboinense var. minor (Bedd.) Bedd., Suppl. Ferns S. Ind. & Brit. Ind. : 19 

1876). 

Peis cation molle var. amboinense sensu Bedd., Handb. Ferns Brit. Ind. : 278 (1883) p.p.—Nephrodium 
amboinense sensu Bedd., Handb. Suppl. : 75(1892) p.p. 

Thelypteris srilankensis Panigrahi in Notes R. bot. Gdn. Edinb. 33: 499 (1975). Type: as for C. 
zeylanica. 

Rhizome short-creeping; fronds up to 60 cm long including stipe but commonly less; stipe to 

9 cm, scaly at the base, elsewhere glabrous as is the rhachis save for scattered hairs in the 

dorsal groove especially distally on the rhachis. Lamina with up to 20 pairs of free subentire 

pinnae beneath the well-marked terminal pinnatifid pinna which may be up to 13 cm long; 

lower 3-6 pairs of pinnae gradually reduced, the lowermost 1-2 cm long; pinnae up to 

9 x 1:2 cm, acuminate, edges crenate, veins 3-5 pairs, 14-2 pairs anastomosing, /ower surface 

glabrous save for minute, gland-like, capitate hairs when young, glabrescent later, upper 

surface with scattered hairs on the costae, elsewhere glabrous. Sori medial, 1-2 pairs of veins 

fertile; indusia glabrous. 


Collections: Thwaites C.P. 3391 (BM: CGE; K (errore 3390); P; PDA). 1887, Wall (P.) 
Kitulgala, 150 m, Sabaragamuwa Province. 
Ceylon and (teste Holttum) Nicobar Islands. 


Christella zeylanica is a very distinct species, differentiated from all other Ceylonese species 
of the genus by its subentire to crenate, nearly glabrous pinnae. In appearance it resembles a 
small form of Sphaerostephanos arbuscula more than the other species of Christella. It is 
related to C. subpubescens, with which it was included by Beddome in his Handbook (as 
Nephrodium molle var. amboinense or N. amboinense). 

Before Fée or Beddome had described this species, Thwaites (Enum. Pl. Zeyl. : 391 (1864)) 
had cited C.P. 3391 and C.P. 3498 as Aspidium extensum Blume (= Amphineuron 
opulentum (Kaulf.) Holttum), gatherings of the Amphineuron being cited under Aspidium 
ochthodes Kunze (loc. cit. : 392). There are two good sheets of C.P. 3391 in the Peradeniya 
Collection, named Nephrodium amboinense Presl, which doubtless represent the original 
gatherings. These are labelled ‘Kittool Galle, Ap. 1855’. Wall, who knew it in this locality, 
followed Thwaites in including the distinct but as yet undescribed Christella papilio with it 
(Cat. Ceylon Ferns : 6 (1873)), using the name Nephrodium amboinense to cover both. A 
year after Fée described N. zeylanicum, Beddome described the same species as N. extensum 
var. minor, subsequently (/oc. cit. 1876) transferring the variety to N. amboinense. Later still he 
dropped the variety, combining Christella zeylanica and C. subpubescens in his Handbook 


40 W. A. SLEDGE 


under the epithet amboinense, at first as a variety of Nephrodium molle and then, in the 
Supplement, as an independent species. Ceylon botanists, however, to whom Christella sub- 
pubescens was unknown, construed Nephrodium amboinense differently, to cover Christella 
zeylanica and C. papilio. The Peradeniya sheets of these species bearing Thwaites’s original 
nomenclature of the Enumeratio are renamed Nephrodium amboinense, and despite the wide 
disparity between them both are, in one instance, mounted on the same sheet. 
Christella papilio remained undescribed until 1899, C.P. 3498 being included by Beddome 
in his Handbook under Nephrodium procurrens and in the Supplement to his Handbook 
under N. molle var. major. 

Though Thwaites cited Uva Province as the locality for C.P. 3391 and C.P. 3498, it was in 
fact the source only of the latter; there is no evidence that Christella zeylanica has ever been 
collected anywhere but about Kitulgala, nor has it been refound there in the present century. 


Hybrids 


Cytological evidence proving the occurrence of hybrids in wild populations of Christella was 
first demonstrated by Manton (in Manton & Sledge, Phil. Trans. R. Soc. Lond. B, 
238 : 127-185 (1954)). That hybridisation is probably of frequent occurrence was suggested 
by the fact that of 18 plants investigated, five of them, from four different stations, gave 
triploid counts or revealed meiotic irregularities. Hybrids within the C. parasitica group 
have also been synthesised experimentally by Panigrahi & Manton (J. Linn. Soc. (Bot.) 
55 : 729-743 (1958)). They have shown that since some important distinguishing characters 
behave as simple dominants in F1 hybrids, such hybrids may bear so close a resemblance to 
one only of the two parent species as to make recognition difficult in the absence of 
cytological evidence. They found dominant characters to be a creeping, as opposed to an 
erect rhizome, non-decrescent versus decrescent frond and presence of subfoliar glands as 
opposed to their absence. Depth of cutting of pinnae, number of anastomosing veins, and 
hair length did not behave as simple dominants or recessives, and hence F1 hybrids were 
intermediate between the parents in these characters. 

Plants with mixed characters are not uncommon in herbaria, and such specimens are best 
regarded as hybrids. In Panigrahi & Manton’s experiments it was found that all the Fl 
hybrids they produced showed a high proportion of abortive spores, and they considered that 
spore sterility would be an effective means of detecting hybridity in a herbarium specimen. 
The converse, however, that good spores disprove hybrid origin, cannot be true since three of 
the hybrid plants listed by Manton (in Manton & Sledge, op. cit. 138) were retained in 
cultivation at Kew, and fronds taken a year or two later by Alston from two of these showed 
sound spores. I believe also that more evidence is required to confirm the dominance of non- 
decrescence, for as Holttum has commented (in Kew Bull. 31:295 (1976)), this is a 
surprising and unexpected finding in a group where non-decrescence is a very much less 
frequent condition than decrescence. 


1. Christella dentata x parasitica 


C. malabariensis sensu Holttum in Kew Bull. 31 : 317 (1976) p.p., non Nephrodium malabariense Fée. 


Frond outline of Christella dentata, but one pair only of reduced and deflexed pinnae; 
pinnae with abundant parasiticus glands below, one pair of veins anastomosing, short-hairy 
beneath, without acicular hairs, some capitate hairs above; indusium hairy. 


Collection: Near Ginigathena, Central Province, 600 m, Dec. 1950, Manton Z 30; Alston 11745 ex 
Trop. Fern House, Kew (BM). 


Since Christella dentata and C. parasitica are the commonest species of Christella in Ceylon and 
India where they often grow together, this is almost certainly the most frequently 
occurring hybrid. The above description is taken from the cultivated plant which shows good 
evidence of each species; it also agrees with Panigrahi & Manton’s statement as to non- 


THELYPTERIDACEAE OF CEYLON 41 


decrescence of the frond being a dominant character, but disagrees with them in having good 
spores. I accept as the same hybrid specimens in which the pinnae are glandular beneath, but 
more than one pair of lower pinnae is shortened. Such plants were included amongst the 
specimens distributed by Kew as Nephrodium molle coll. Stocks & Law and these were in my 
opinion wrongly included by Holttum in Fée’s N. malabariense. A full discussion of this 
matter is given in my paper in Kew Bull. 34 : 77-81 (1979). 

It is probable that several forms of this hybrid exist. From Lady Horton’s Walk, Kandy, I 
have two gatherings (509 and 1357) which may be hybrids of the same parentage, though 
they are very different from the Ginigathena plant. In both gatherings the fronds are set 
2-3 cm apart on a wide-creeping rhizome, and both have abortive sporangia and spores. 
These are Nephrodium procurrens sensu Baker (Syn. Fil. : 290 (1867)) and of Beddome’s 
Handbook (278 (1883)), which later authors have referred either to Christella dentata (e.g. 
Ching in Bull. Fan mem Inst. Biol. (Bot.) 8 : 207 (1938)) or to C. parasitica (e.g. Holttum in 
Kew Bull. 31 : 309 (1976)). If the long-creeping rhizome and sterility denote a hybrid origin, 
then C. dentata x parasitica would seem to be the only possible parentage, though the 
absence of glands, long acicular hairs or capitate hairs or any other character indicative of 
parasitica makes such an identification doubtful. 


2. Christella hispidula x parasitica 


Frond outline of Christella parasitica with lowest pair only of pinnae somewhat shortened; 
one pair of veins anastomosing, the second pair reaching margins well above base of sinus; 
pinnae eglandular beneath rather thinly hairy on costae and veins and on indusia, without 
long acicular hairs. Sporangia mostly abortive; no good spores. 


Collection: Lady Horton’s Walk, Kandy, 600 m, Dec. 1950, Manton P43; Alston 11742 ex Trop. Fern 
House, Kew (BM). 


This plant was a triploid with N pairs and N singles. The frond taken by Alston from the 
cultivated stock agrees well with Panigrahi & Manton’s silhouette (/oc. cit. fig. 4) of a 
synthesised hybrid of the same parentage, save that the lowermost pair of pinnae is 
somewhat shorter than the next pair. The absence of subfoliar glands in the wild hybrid is 
presumably due to the Christella parasitica parent, in this instance being an eglandular form 
of the species. Both parents and C. dentata are frequent on the wooded hillside of Lady 
Horton’s Walk but the venation and triploid cytology would appear to rule out C. dentata as 
the second parent. 


3. Christella meeboldii x parasitica 


Rhizome ‘sub-erect to short-creeping’; fronds decrescent with three to six pairs of 
progressively reduced pinnae; pinnae cut less than half way to costa, short-hairy on costa and 
veins above and below, without acicular hairs but with abundant parasiticus glands beneath, 
two pairs of veins anastomosing; indusia glabrous or nearly so; spores + verrucose or shortly 
ridged. 


Collection: Hangiliella beyond Welimada on Badulla road from Nuwara Eliya, c. 900 m, 29 Dec. 1950, 
Manton P. 307; Alston 11737 ex Trop. Fern House, Kew (BM). 


The mixed characters of this plant seem only explicable on the assumption of the above 
parentage, and that non-decrescence of the frond is not a dominant character in this hybrid 
combination. Typical C. meeboldii and C. parasitica were both collected here by Ballard. 


XII. PNEUMATOPTERIS Nakai 


in Bot. Mag. Tokyo 47: 179 (1933), emend. Holttum in Blumea 19:42 (excl. Pseudocyclosorus) 
(1971); op. cit. 21 : 293 (1973). 


42 W. A. SLEDGE 


Caudex usually erect, rarely creeping; rhizome scales broad, thin, with marginal hairs; fronds 
pinnate, usually large, decrescent, with shallowly to deeply lobed pinnae, the lobes with 
cartilaginous margins; usually several pairs of basal pinnae reduced either abruptly or 
gradually; aerophores on reduced and lower pinnae distinct, + swollen; stipe and lamina 
never conspicuously hairy; veins in most species anastomosing, free in a few species; lamina 
between veins + pustular when dry, sessile spherical glands never present; sori usually 
indusiate, sporangia often bearing short, club-shaped, glandular hairs, stalks with a 2-4 
celled hair with enlarged terminal cell; spores pale with many small + quadrate wings of 
irregular shape, hence spinulose in aspect. n = 36. 

About 75 species; mainly in Malesia with a few species in Africa and the Mascarene 
Islands; mainland Asia from southern China southwards throughout Malesia to Australia 
(northern Queensland) and New Zealand, and in the Pacific Islands from Fiji and Samoa to 
Hawaii. 


1. Pneumatopteris truncata (Poiret) Holttum 


in Blumea 21: 314 (1973).—Polypodium truncatum Poiret, Encycl. Meth. 5: 534 (1804). Type: Brazil 
[no collector’s name] (P).—Cyclosorus truncatus (Poiret) Farwell, Am. Midl. Nat. 12 : 250 (1931). 
Ching in Bull. Fan meml Inst. Biol. (Bot.) 8 : 216 (1938). 

Nephrodium truncatum sensu Bedd., Handb. Ferns Brit. India: 280 (1883), non (Gaudich.) C. Presl. 

Aspidium abortivum Blume, Enum. Pl. Jay. : 154 (1828). Type: Java, Blume (L). 

Aspidium abruptum Blume, Enum. Pl. Jav.: 154 (1828). Type: Java, Kuhl & van Hasselt (L).— 
Nephrodium abruptum (Blume) J.Sm. in Hooker’s J. Bot. 3: 411 (1841). Hook., Spec. Fil. 4:77 
(1862) p.p. et excl. t.241 B. 

Aspidium eusorum Thwaites, Enum. Pl. Zeyl. : 391 (1864). Type: Ceylon, Thwaites C.P. 3064 (K).— 
Nephrodium eusorum (Thwaites) Bedd., Ferns Brit. India : t.130 (1866) [Illustration of Beddome 
(K)]. 

Rhizome erect, fronds tufted, from 90-240 cm long; stipe and rhachis pale, puberulous at 

first, becoming glabrous; lamina 60-120 x 30-60 cm, bipinnatifid with 20-30 pairs of pinnae 

and up to 8 pairs of abruptly reduced lower ones forming widely spaced auricles along the 
stipe; pinnae up to 30 x 3 cm in large fronds, 15 x 1:5—2 cm in small ones, apex acuminate, 
base truncate in the upper pinnae becoming contracted below and shallowly lobed in the 
lower pinnae, margins lobed up to } way to the costa, the lobes oblong and slightly 

crenated about the subtruncate apices; veins up to 10 pairs, strongly raised beneath, 14-24 

pairs anastomosing; upper surface of costae sparsely hairy otherwise surfaces glabrous above 

and below but with pustular swellings beneath when dry; sori medial, forming two parallel rows, 
the lower ones not divergent; indusia glabrous, small, evanescent. 


Collections: Thwaites C.P. 3064 (CGE; K; P; PDA). Hantane Range, in forests, July 1844, Gardner 
1104 (BM; CGE; K). Kadugannawa, shady forest, Oct. 1846, Gardner 1252 (CGE, K). Oodawella, 
1870, leg. Randall in herb. Rawson 3244 (BM). Gallebodde, Central Province, by stream in jungle, 
600 m, 26 Jan. 1954, Sledge 1046 (BM). Between Gilimale and Carney, Ratnapura District, jungle, 
150 m, 9 March 1954, Sledge 1250 (BM). Sinha Raja Forest above Beverley Estate, Deniyaya, 900 m, 
12 March 1954, Sledge 1277 (BM). Lady Horton’s Walk, Kandy, 600 m, 24 March 1954, Sledge 1355 
(BM). Unlocalised: Beddome (K). 1887, Wall (P). Mrs Chevalier (BM). Macrae (CGE). Bradford 
(CGE). Walker (K). 
In forests of the interior to 1500 m. 

Also in southern and north-east India, southern China, western Malesia and the 
Philippine Islands. 


A large fern, in facies most resembling Pronephrium articulatum and confused with that 
species by some early botanists. It is readily distinguished by its decrescent fronds with 1-2 
pairs of anastomosing veins in the glabrous pinnae, as opposed to the non-decrescent, 
glandular fronds of P. articulatum, with at least 3 and commonly 4-6 pairs of anastomosing 
veins. 


THELYPTERIDACEAE OF CEYLON 43 
XIV. SPHAEROSTEPHANOS J.Sm. 


in Hook., Gen. Fil. : t.24 (1839). Holttum in Blumea 19 : 39 (1971). 


Caudex erect to long-creeping; fronds decrescent with varying number of much-reduced 
basal pinnae; pinnae pinnatifid to shallowly lobed; aerophores at base of pinnae often 
swollen; veins anastomosing or rarely free, surfaces always + hairy and sessile spherical 
glands commonly present on lower or both surfaces; sori round or in a few species + 
elongate, usually indusiate; indusia often hairy and/or glandular; sporangia usually bearing 
spherical glands or setae near the annulus, the stalks bearing a multicellular hair with 
swollen end-cell; spores light brown, spinulose or bearing many small translucent wings. 
n= 36. 

About 150 species: mainly in New Guinea and the Philippine Islands with a few species in 
Madagascar, Mascarene Islands, southern India and Ceylon, Burma, southern China and 
throughout Malesia eastwards across the Pacific to Tahiti. Outlying stations in S40 Tomé, 
east Africa, New Caledonia and Australia. 


1. Sphaerostephanos arbuscula (Willd.) Holttum 


in Jl. S. Afr. Bot. 40: 164 (1974).—Aspidium arbuscula Willd. in L., Sp. Pl. 4th ed., 5: 233 (1810). 
Type: Mauritius. Herb. Willd. 19,763 (B).—Nephrodium arbuscula (Willd.) Desvaux in Mém. Soc. 
linn. Paris 6: 253 (1827). Beddome. Handb. Ferns Brit. Ind. :276, fig. 142 (1883).—Dryopteris 
arbuscula (Willd.) Kuntze, Rev. Gen. Pl. 2: 812 (1891).—Cyclosorus arbuscula (Willd.) Ching in 
Bull. Fan mem inst. Biol. (Bot.) 8 : 194 (1938).—Thelypteris arbuscula (Willd.) K. Iwats. in Acta 
phytotax. geobot. Kyoto 21 : 170 (1965S). 


Rhizome erect, often emergent and trunk-like; stipe and rhachis grey-pubescent, the stipe 
short and scaly at the base; fronds tufted, commonly 30-60 cm but sometimes reaching | m 
or more, 10-20cm wide, oblong-lanceolate in outline and narrowing towards both 
extremities, pinnate with numerous pairs of narrowly oblong, acuminate pinnae often 
auricled at the acroscopic base, margins crenate or serrate, up to 10 pairs and sometimes 
more of the lower pinnae gradually reduced to deltoid auricles and extending nearly or quite 
to the base of the stipe; veins 4-6 pairs the lowermost 1-1} pairs anastomosing and forming 
an excurrent nerve, the next pair running to the long sinus membrane; texture herbaceous; 
upper surface of pinnae glabrous with age save on the costa, /ower surface hairy on the costa, 
costules and veins and dotted with spherical, sessile, yellow glands; sori medial on the veins, 
indusium glandular with few or many hairs; sporangia with yellow glands near the annulus. 


Collections: Thwaites C.P. 1359 (BM; CGE; K; P; PDA). Rambodde, shady banks, June 1845, Gardner 
1109 (CGE; K). Kadugannawa, moist woods, Oct. 1846, Gardner 1251 (CGE). Adam’s Peak, 14 Feb. 
1908, Matthew (K). Kotmalee, 1847, Fortescue (CGE). Badulla, Freeman 264, 265 (BM). 1870, 1871, 
Oodawella, Randall in herb. Rawson 3220 (BM). Kandy, 1854, Bradford (BM; CGE; P). Hunnasgiriya, 
18 Jan. 1954, Schmid 951 (BM). Galaha 940 m, 22 Jan. 1954, Schmid 1036 (BM). Deniyaya, 550 m, 5 
Feb. 1954, Schmid 1145 (BM). Balangoda-Rassagala road, Ratnapura Dist., in lowland forest, c. 750 
m, 16 Nov. 1976, Faden 76/306 (K). Above Pinnawala on Balangoda road, rocky creek in forest, 19 
March 1968, Comanor 1090 (US). Corbet’s Gap, 1200-1300 m, 9 Dec. 1950, Sledge 553, 557 (BM). 
Same locality, 1200 m, 7 Jan. 1951, Sledge 854 (BM). Ramboda Pass, 1560 m, 17 Dec. 1950, Sledge 
660 (BM). Hunnasgiriya, open ground by stream, 870 m, 16 Jan. 1954, Sledge 969 (BM). Amba- 
gamuwa, marsh in jungle, 570 m, 19 Jan. 1954, Sledge 993 (BM). Unlocalised: Walker (K). Robinson 
150 (K). Geo. Wall (P; PDA). Ferguson (PDA; US 816397). 


Common on banks of streams and in wet ground in forests in the interior from 550-1550 m. 


Southern India, Ceylon, Mascarene Islands, Madagascar and, as subsp. africanus 
Holttum, Tanzania and Kenya. 


This species and Sphaerostephanos unitus are two of the most frequently occurring 
thelypteroid ferns in the hill country of Ceylon. Several other species with anastomosing 
veins have coloured glands on or between the veins on the under surface of the pinnae, but 


44 Ww. A. SLEDGE 


the species of Sphaerostephanos are the only ones in which this character is combined with 
decrescent fronds. 


2. Sphaerostephanos subtruncatus (Bory) Holttum 


in Kew Bull. 26 : 80 (1971).—Polypodium subtruncatum Bory in Bélanger, Voy. Ind. Or. (Bot.) 2 : 32 
(1833). Type: India, Madura, Mts de Dendigall, Bélanger (P). 

Dryopteris mauritiana var. gardineri Christensen in Trans. Linn. Soc. (Bot.) Il, 7: 413 (1912). Type: 
Seychelles, J. S. Gardiner (K). 


Rhizome erect, fronds tufted, up to 2 m tall; stipe 40-60 cm long, dark coloured, tetragonal, 
scaly near the base, elsewhere glabrous or nearly so, bearing about 15 pairs of contracted 
pinnae, the lower ones reduced to mere papillate protuberances; lamina elliptic-lanceolate, 
75-120 cm with 30 or more pairs of pinnae, the transition between normal and reduced 
pinnae abrupt; pinnae up to 16 x 1:5 cm sessile, bearing a brown aerophore at junction with 
rhachis, apex acuminate, margins lobed half way to costa, lobes blunt, slightly falcate; veins 
5-8 pairs in each lobe, 1} pairs anastomosing }-1 pair to base of sinus; costae and costules 
hairy above and below, scattered hairs on the veins and numerous sessile, spherical, yellow 
glands present throughout lower surface; texture firm herbaceous; sori medial, indusia 
glabrous or with a few short hairs. 


Collections: Thwaites C.P. 714 (2 sheets), sub Nephrodium molle (PDA). Hillside above Potupitiya, 
Ratnapura District, in degraded forest, 450 m, 4 Dec. 1976, Faden 76/481 (K). 
Rare. 

Also in south-west India and the Seychelles. 


Sphaerostephanos subtruncatus differs from S. unitus in its erect rhizome and softer, more 
deeply lobed, pinnae which are much less hairy beneath and without prominently raised 
veins. The Peradeniya examples agree closely with specimens from south-west India, save 
for their subglabrous stipes. The species appears not to have been re-collected in Ceylon 
until Faden’s 1976 gathering; the label on his specimen states that only one plant was seen. 

Holttum cites four collections from India, and Christensen one from the Seychelles. A 
specimen from India at Genéve, received from Kew and named as Nephrodium molle, is 
another example of Sphaerostephanos subtruncatus. A third sheet of Thwaites C.P. 714 at 
Peradeniya is Christella hispidula (Decaisne) Holttum, as is a sheet bearing the same number 
at Kew. Three sheets of Thwaites C.P. 714 at P and others at CGE and BM all represent C. 
dentata (Forssk.) Brownsey & Jermy. 


3. Sphaerostephanos unitus (L.) Holttum 


in JL. S. Afr. Bot. 40: 165 (1974).—Polypodium unitum L., Syst. Nat. 10th ed., 2 : 1326 (1759), excl. 
syn. Type: no locality (LINN).—Aspidium unitum (L.) Swartz in J. Bot. Gottingen 1800 (2): 32 
(1801) nomen tantum. Sensu Thwaites, Enum. Pl. Zeyl. : 391 (1864), non Mettenius—Dryopteris 
unita (L.) Kuntze, Rev. Gen. Pl. 2: 811 (1891).—Cyclosorus unitus (L.) Ching, in Bull. Fan meml 
Inst. Biol. (Bot.) 8 : 192 (1938). Holttum, Rev. Fl. Malaya 2 : 260, fig. 147 (1955). 

Aspidium cucullatum Blume, Enum. Pl. Jav.: 151 (1828). Type: Java (L).—Nephrodium cucullatum 
(Blume) Bedd., Ferns S. Ind. : t.88 (1863). Baker in Hook. & Baker, Syn. Fil. : 290 (1867). Bedd., 
Handb. Ferns Brit. Ind. : 270, fig. 138 (1883).. 


Rhizome wide-creeping, apical region covered with narrow, brown scales; stipes 30-50 cm 
long to lowest unreduced pinnae, scaly at the base elsewhere villose with short appressed 
hairs but becoming glabrescent with age; rhachis densely hairy; lamina 40-60 cm long, 
pinnate with 15-25 (30) pairs of pinnae excluding several pairs of abruptly reduced, auricle- 
like lower pinnae, apex of frond pinna-like; pinnae up to 15 cm long, 0:75—1°5 cm wide, apex 
acuminate, base broadly cuneate, margins cut down about one third into rounded, acute, 
cucullate lobes; veins 8-10 pairs per lobe, thick and raised beneath, 1-11 pairs anastomosing 


THELYPTERIDACEAE OF CEYLON 45 


the next two pairs running to the long hyaline sinus membrane; upper surfaces of pinnae 
glabrous save for the strigose-hairy costae, /Jower surfaces of costae, costules and veins rather 
densely clothed with pale, stiff hairs, the veins and lamina surface also dotted with small 
spherical sessile orange or yellow glands; texture of frond subcoriaceous; sori supramedial on 
the veins which are generally all soriferous; indusium firm, persistent, usually eglandular and 
more or less hairy; sporangia glandular. 


Collections: Thwaites C.P. 973 (BM; CGE; K; P; PDA). Peradeniya, 24 May 1915, Petch (PDA). 
Kotmalee, 1847, Fortescue (CGE). Badulla, Freeman 263 (BM). Common at Kandy, Mrs Chevalier 
(BM). Same locality, 1854, no collector’s name (CGE). Same locality, 1868, 1871, Randall in herb. 
Rawson 3220 (BM). Galle, woods, May 1860, Dubuc (E). Point de Galle, 1875, Lischke (P). Caltura, 
woods, May 1820, Leschenault (P). Hunnasgiriya, 900 m, 18 Jan. 1954, Schmid 954 (BM). Rajawaka, 
Ratnapura District, moist roadside bank, 460 m, 30 Dec. 1976, Faden 76/660 (K). Between Hakgala 
and Ambawela, roadside 1650 m, Dec. 1950, Sledge 716, 722 (BM). Pussalamankada, between Kandy 
and Maturata, open ground above stream, 540 m, 18 Jan. 1954, Sledge 982 (BM). Unlocalised: 1819, 
Moon (BM). Ferguson (PDA; US 816403). 


Common on moist roadside banks and open ground near streams in the west and centre, up 
to 1650 m. 

East Africa, Madagascar and Mascarene Islands, Seychelles, southern India, Assam and 
Burma to Vietnam, throughout Malesia to New Guinea and the Philippine Islands, New 
Caledonia, Fiji and Samoa. 


XV. PRONEPHRIUM Pres! 


Epimel. Bot. : 258 (1851). Holttum in Blumea 19 : 34 (1971); 20 : 105 (1972). 


Caudex suberect to long-creeping; fronds simply pinnate (rarely simple) with subentire 
pinnae and terminal pinna-like segment, the basal pinnae not reduced; most pairs of veins 
anastomosing to form united excurrent vein between costules; lower surface of pinnae often 
pustular when dry, acicular or hooked hairs frequent on one or both surfaces; sori indusiate, 
or exindusiate and then spreading along veins, sporangia often bearing hairs or glands near 
annulus; spores with a continuous wing and a few cross-wings. n = 36. 

About 60 species; India and Ceylon, southern China southwards throughout Malesia; 
north-eastern Australia and the Pacific Islands. 

Holttum (loc. cit.) divided this genus into three Sections. The smallest of these, Sect. 
Grypothrix Holttum, is distinguished by the presence of hooked hairs on the pinnae and 
sporangia; two of the four Ceylonese species of Pronephrium belong to this group. They are 
widely different in appearance from the other two species which belong to the two other 
generic Sections. 


1. Pronephrium articulatum (Houlst. & Moore) Holttum 


in Blumea 20: 116 (1972).—Nephrodium articulatum Houlst. & Moore, Gdnrs’ Mag. Bot. Hort. Flor. 
: Nat. Sci. : 293 (1851). Type: cult. Hort. Bot. Kew ex Ceylon, 1845, Gardner 1104 bis (BM; CGE; 
el articulatum (Houlst. & Moore) Lowe, Ferns Brit. & Exot. 6: t.49 (1857), non Swartz 
(1801). 
Nephrodium abruptum sensu Hook., Spec. Fil. 4 : 77-78, t.241B (1862) p.p., non Aspidium abruptum 
Blume (1828). Bedd., Ferns S. Ind. : 31, t.86 (1863), non Aspidium abruptum Blume.—Aspidium 
abruptum sensu Thwaites, Enum. Pl. Zeyl. : 391 (1864), non Blume (1828). 
_ Nephrodium pennigerum Baker in Hook. & Baker, Syn. Fil. 2 ed. : 292 (1874) p.p. C. B. Clarke in 
Trans. Linn. Soc. (Bot.) Il, 1: 532 (1880), quoad plantae Zeylanicae. Bedd., Handb. Ferns Brit. 
| Ind. : 277 (1883) p.p.; Handb. Suppl. : 73 (1892). 
_Nephrodium glandulosum var. laete-strigosum C. B. Clarke in Trans. Linn. Soc. (Bot.) Il, 1 : 532, t.74 
fig. 2 (1880). Type: [Bangladesh,] Chittagong, Clarke 19900 (K). 
Dryopteris megaphylla (Mettenius) Christensen, Index Fil. : 277 (1906) p.p. quoad plantae Zeylanicae. 
Dryopteris indica van Rosenb., Malay. Ferns : 224 (1909). 
Cyclosorus laete-strigosus (Clarke) Ching in Bull. Fan mem Inst. Biol. (Bot.) 8 : 227 (1938). 


46 W. A. SLEDGE 


Rhizome short-creeping, stipes up to 50 cm long, sometimes longer, with scattered broad 
thin brown scales especially in the lower part and sparsely hairy throughout as is the rhachis; 
lamina 50-80 (100) cm, pinnate with up to 20 pairs of patent pinnae below the terminal 
pinna; /ower pinnae not reduced, largest pinnae up to 20 x 3 cm apex acuminate, base of 
upper ones truncate to broadly cuneate, lower ones more narrowly cuneate, margins 
shallowly lobed to about | way to the costa, the lobes rounded or subacute; veins up to 12 
pairs about 3-5 pairs anastomosing the next 2-3 running into the sinus membrane, veins in 
the terminal pinna often forking and anastomosing; costae and veins sparsely appressed- 
hairy above and beneath with short stiff hairs on the lamina surface (sometimes rather 
plentiful on the upper surface), the Jower surface with scattered small, sessile ochre-coloured 
glands on and between the veins; fertile fronds contracted with narrowed pinnae; texture 
herbaceous; sori medial, indusia glabrous; spores spinulose. 


Collections: Thwaites C.P. 3271 (BM; CGE; K; P; PDA: Haldingmulle, April 1856; Wattegodde 1856). 
Hantane range, in forests, Aug. 1844, Gardner 1104 (bis) (BM; CGE). Lady Horton’s Walk, Kandy, in 
forest, 600 m, Feb., March 1954, Sledge 1144, 1363 (BM). Unlocalised: 1884, Wall (PDA). W. 
Robinson C. 155 (K). 


Forests of Central Province from 600-1200 m. 
Also in India, Bangladesh, Burma, northern Thailand and western China. 


Holttum’s description refers to the rootstock as being erect though Houlston & Moore’s (/oc. 
cit.) original description, described from cultivated plants at Kew, alludes to a ‘thick creeping 
rhizome.’ Beddome’s description in the Handbook states “caudex erect’, but this is corrected 
in the Supplement to ‘often, if not always, more or less creeping.’ My no. 1363 has a creeping 
rhizome, though this is not evident in my no. 1144. 

This species was described from Ceylonese plants first found by Gardner in 1844 and sent 
to Kew by him in 1845. It has probably been a rather rare species confined to forests at a 
moderate elevation in the Central Province, and since these have mostly now been cleared 
for plantations, it will certainly be rarer now. My collections from Kandy are the only recent 
gatherings. 


2. Pronephrium gardneri Holttum 


in Kew Bull. 26: 81 (1971). Type: Ceylon, Gardner 1137 (K, holotype). [non Gardner 1135, Holttum 
loc. cit.] 
3063 (K)].—Nephrodium urophyllum sensu Bedd., Ferns S. India Suppl. : 18 (1876); Handb. Ferns 
Brit. Ind. : 274 (1883) p.p. 
Polypodium granulosum sensu Thwaites, Enum. Pl. Zeyl. : 394 (1864) p.p., non Presl. 
Stipes 50 cm or more long, stramineous, glabrous save in the groove. Fronds 70 cm or more 
long with 10-12 pairs of pinnae, the lowermost up to 25 x 3°5 (4) cm, narrowly oblong, 
shortly acuminate at the apex, base truncate or broadly cuneate (save the lower ones which 
are narrowed towards the base), margins coarsely crenate, shortly ciliate; costules 
4:5-5°5 mm apart; veins 10-11 pairs, slender, slightly raised, 7 pairs anastomosing (the 
excurrent nerves often free in lower parts), two pairs joining the sinus membrane, three or 
four veins reaching the margin. Lower surface of rhachis glabrous, of the costae minutely 
pilose, lamina minutely pustulose; upper surface of rhachis and costae shortly pilose. Sori 
medial, round, exindusiate; sporangia smooth, spores with a translucent wing with minutely 
erose margin and a few transverse wings. 


Collections: Hantane Range, in forests, Aug. 1844, Gardner 1137 (CGE; K, holotype). Thwaites C.P. 
3063 (P; PDA: Hantane, Jan. 1854). Thwaites C.P. 3063 (K). 


Forest near Kandy. 
Endemic to Ceylon. 


THELYPTERIDACEAE OF CEYLON 47 


Thwaites was surely wrong in stating that this fern was ‘not uncommon’ in the forests of the 
Central Province. Very few collections exist (only five sheets having been located) and all are 
from the vicinity of Oodawella in the Hantane range near Kandy. No forest now remains on 
these hills, and since no specimens have been found elsewhere, it is probable that this species 
is extinct. 


3. Pronephrium triphyllum (Swartz) Holttum 


in Blumea 20 : 122 (1972).—Meniscium triphyllum Swartz in J. Bot. Gottingen 1800 (2): 16 (1801): 
Hook & Baker Syn. Fil. 2nd ed. : 391 (1874). Bedd., Handb. Ferns Brit Ind. : 397, t.231 (1883). Type: 
No locality or collector (S-PA).—Abacopteris triphylla (Swartz) Ching in Bull. Fan mem. Inst. Biol. 
(Bot.) 8 : 241 (1938). 

Rhizome long-creeping, clothed near the apex with linear, brown, ciliate scales; stipes 

sulcate, hairy, 30 cm or more long in fertile fronds, sterile fronds with shorter stipes; /amina 

trifoliate or with five pinnae, the terminal pinna largest, oblong-lanceolate, apex acute or 
acuminate, base rounded, margins entire or sinuate, glabrous above save on the costa, with 
hooked hairs on the costa and veins beneath; texture herbaceous; terminal pinna of sterile 
fronds up to 15 cm long, 2-3} cm broad, of fertile fronds usually narrower; veins normally 

8-10 pairs anastomosing at a wide angle, the excurrent vein often free though usually united 

to the next pair of veins above, in the fertile pinnae the veins almost at right angles to the 

costules, the exindusiate sori occupying almost the whole length of the veins; sporangia 
setose with hooked hairs, spores spinulose. 


Collections: Thwaites C.P. 1293 (CGE; K; P; PDA: Pasdun Korle, Dec. 1845, Gardner; Ambagamuwa, 
Nov. 1854; Hunnasgiriya, April 1857). Pussilawa, 1847, E. Tennents (CGE), Ferguson (PDA). By 
track from Palawatta to Pahale Hewissa, on sandy ground in shade, 45 m, 20 Jan. 1951, Sledge 886 
(BM). Gongalla Hill, + 1000 m, 11 March 1954, Sledge 1261 (BM). Near Hiniduma, Southern 
Province, shady bank by path, 3 April 1954, Sledge 1393 (BM). 


Damp, shady places at low to moderate elevations. 
India, Ceylon, Burma, Thailand, Malaysia, China, Japan, Taiwan, the Philippine Islands 
and Australia (Queensland). 


Beddome’s Meniscium parishii, described from Burma, has four or five pairs of free pinnae 
beneath the terminal pinna. A specimen of Griffith’s from north-east India has six pairs. 
Beddome reduced his species to a variety of M. triphyllum in his Handbook and Holttum 
also treated it as a variety in his Ferns of Malaya. However, in his account of the genus 
Pronephrium, Holttum (Blumea, loc. cit.) raises parishii to specific rank and alters the 
circumscription to include all plants with more than one pair of pinnae, thus including 
Ceylon in its distribution. Yet fully mature fertile fronds bearing three or five pinnae may be 
found attached to the same rhizome. It is also improbable that if populations of plants with 
five pinnae existed in Ceylon that these would have been overlooked by collectors. I have 
seen no specimens from Ceylon which match Beddome’s illustration (Ferns Brit. Ind. t.184 
(1866)) of his Meniscium parishii. Whatever the status of plants from Burma and northern 
poe I see no grounds for the recognition, even at varietal level, of another taxon from 
eylon. 


4. Pronephrium thwaitesii (Hook.) Holttum 


in Blumea 20 : 122 (1972).—Meniscium thwaitesii Hook., Fil. Exot. sub t.83 (1859). Thwaites, Enum. 
Pl. Zeyl.:382 (1864). Bedd., Handb. Ferns Brit. Ind.:399, fig. 232 (1883). Type: Ceylon, 
Thwaites C.P. 3145 (K).—Abacopteris thwaitesii (Hook.) Ching in Bull. Fan meml Inst. 
Biol. (Bot.) 8 : 243 (1938). 

Rhizome wide-creeping; stipes up to 30 cm long, nearly naked save at the base; rhachis 

hairy; lamina subdeltoid in outline 15-25 x 10-15 cm, pinnate with 2-4 pairs of free pinnae 

beneath the apical pinna which is pinnate below with adnate segments, becoming pinnatifid 


48 W.A.SLEDGE | 


then lobed then crenate distally. Lowest pair of pinnae the largest, 5-10 x 1-14 cm, subsessile 
or shortly stalked, apex acuminate, margins crenate or shallowly lobed, divisions of the 
apical pinna entire, narrower and blunter and often falcately curved; surface of pinnae 
glabrous above save on the costa, hairy with hooked hairs on the veins beneath, areoles 4—5 
in the lower pinnae between the costa and margin; sori exindusiate, those on lower veins 
elongate, on distal veins circular; sporangia setose with hooked hairs. 


Collection: Thwaites C.P. 3145 (CGE; K, holotype; P; PDA). 


Forests of Central Province from 900-1200 m. 
Also in southern India (Nilgiris Hills). 


Information given on the three sheets in PDA indicates that Pronephrium thwaitesii has 
been collected in a small number of widely separated localities in the Central Province, viz. 
Udapusselawa (April 1854), Bogawantalawa (Nov. 1855) and at Kotagalla and Lagalla in 
Matale East. No collections have been made however in the last 100 years, either in Ceylon 
or southern India. 

Holttum (/oc. cit.) has suggested that this is ‘probably a hybrid between P. triphyllum and a 
species with lobed pinnae.’ Elsewhere he (Blumea 19 : 36 (1971)) suggested Trigonospora 
ciliata as the other possible parent. I see more to refute than to support this suggestion. No 
station is known in which Pronephrium triphyllum and P. thwaitesii occur together. The 
second parent could scarcely be either P. articulatum or P. gardneri, neither of which has 
ever been collected in any of the areas from which P. thwaitesii has been recorded. If 
hybridisation involved a more distantly related species belonging to another thelypteroid 
group, as is implied by Holttum, this would probably be reflected in spore sterility and 
malformed sporangia; yet the sporangia and spores are well formed in the specimens I have 
examined. 

Wall (Cat. Ceylon Ferns, notes: 7 (1873)) was the first to recognise that the entire- 
margined, blunt-ended and falcately-curved distal ‘pinnae’ really belong to the apical pinna, 
and represent its lower, fully pinnated portion, the only true pinnae being the 2-4 pairs of 
stalked, crenate-margined and acuminate basal pairs. He also records the important field 
observation that in East Matale (the only region in which Pronephrium triphyllum is known 
to occur in the same area as P. thwaitesii) ‘the least divided forms of M. Thwaitesii are 
found ...on the same range of mountains, but on the opposite side of the range, where the 
most divided forms of M. Triphyllum grow.’ The inference is that the boundaries between P. 
triphyllum, P. parishii and P. thwaitesii may be less clearly defined than has been supposed. 
The relationships between the three taxa need reinvestigation in the field rather than the 
herbarium. 


[Pronephrium nudatum (Roxb.) Holttum in Blumea 20: 111 (1972) (Nephrodium moulmeinense 
Bedd., Handb. Ferns Brit. Ind. : 275 (1883)) is not a Ceylon fern. Single specimens at CGE, K & PDA 
from Ceylon were doubtless of botanical garden origin.] 


XVI. STEGNOGRAMMA Blume 


Enum. Pl. Jay. : 172 (1828) emend. K. Iwats. in Acta phytotax. geobot. Kyoto 19 : 112-126 (1963). 


Caudex ascending or short-creeping; stipe and rhachis hairy with unicellular or septate hairs; 
fronds pinnate with subentire or pinnatifid pinnae, the basal pinnae not or little reduced, the 
upper ones coadnate at base; aerophores lacking; veins free or with goniopteroid 
anastomosis; surfaces of pinnae hairy, lacking spherical glands; sori exindusiate, elongated 
along the veins, sporangia setose; spores finely spinulose or with many small wings. n = 36. 

About 12 species, from Spain, Macaronesia, tropical and southern Africa eastwards across 
the warmer parts of east Asia to Japan and south to the Philippine Islands and Indonesia; a 
few in tropical America. 


} 
} 


THELYPTERIDACEAE OF CEYLON 49 


1. Stegnogramma pozoi (Lag.) K. Iwats. var. petiolata (Ching) Sledge, stat. nov. 
Leptogramma petiolata Ching in Acta phytotax. sin. 8 : 319 (1963). Type: Ceylon, G. Wail. 
Grammitis totta sensu Thwaites, Enum. Pl. Zeyl. : 382 (1864), non Presl.—Leptogramme totta sensu 
Bedd., Handb. Ferns Brit. Ind. : 377 p.p. excl. fig. 215 (1883). 
Rhizome short-creeping clothed with lanceolate, brown, ciliate scales. Fronds 20-60 cm, 
stipe shorter than lamina bearing scattered brown scales below and a mixture of long and 
short, spreading, white, acicular hairs throughout, rhachis similarly hairy. Lamina 
lanceolate to narrowly oblong-lanceolate, 4-15 cm wide, pinnate with up to 9 pairs of free 
pinnae below the broadly adnate distal ones, several pairs of lower pinnae subsessile or very 
shortly stalked, the lowermost pair sometimes reduced. Pinnae 2-6 (10) x 1-1:75 cm, 
truncate at the base and usually rather suddenly narrowed to a blunt or subacute apex, rarely 
attenuate, lobed less than half way to the costa with the lobes blunt and rounded or only 
crenate-lobate in small fronds, bearing a mixture of some long and many short hairs on the 
costa and veins on both surfaces and often minute hairs on the lamina surface; veins in the 
segments simple, free or rarely with occasional anastomoses; texture herbaceous. Sori linear, 
medial, exindusiate with setose sporangia. 
Collections: Thwaites C.P. 1292 (BM; CGE; K; P). Newara Eliya, in woods, 1800 m, Sept. 1844, 
Gardner 1071 (BM; CGE; P). Same locality: April 1899, Gamble 27567 (K). Moon Plains near Newara 
Eliya, 1800 m, 23 Dec. 1950, Ballard 1201 (K). By track from Pattipola to Horton Plains, 1800 m, 20 
Dec. 1950, Sledge 667 (BM), Horton Plains, by stream in shady gully, 2040 m, 19 Dec. 1950, Sledge 
687 (BM). Horton Plains, 2070 m, March 1954, Schmid 1397 (BM). Same locality, on roadside banks 
in forest, c.2090 m, 15 Nov. 1976, Faden 76/285 (K). Ramboda Pass-Maturata track, in shady forest, 
1940 m, 17 March 1954, Sledge 1305, 1316 (BM). Hakgala, 1800 m, 20 March 1954, Sledge 1340 
(BM). Unlocalised: Walker (K). Bradford 481 (P). May 1906, Matthew (K). Mrs Chevalier (BM). 
Freeman 334A, 335B, 336C (BM). 


In forests about Newara Eliya, above 1750 m. 
Stegnogramma pozoi s. lat. is recorded from Spain, Madeira & west Africa to China and 
Japan; var. petiolata from Ceylon and Java. 


Ching’s description of Leptogramma petiolata is based on two collections from Ceylon; the 
type is said to be based on a collection by ‘Wallich’ and the other gathering is cited as 
Thwaites 481. The former is an error for G. Wall; the latter presumably refers to Bradford 
481 of which there are three sheets from Ceylon in P. Thwaites C.P. 481 is a species of 
Viscum. Although Ching’s description appears to be based on only two sheets, more ample 
material confirms that Ceylon plants differ from south Indian plants in their generally 
smaller stature, their shorter (3-4 cm) pinnae, which are proportionately broader and more 
abruptly narrowed into a blunt or subacute apex (rarely acuminate as described by Ching), 
and in having many pairs of pinnae free from the rhachis and shortly stalked—not ‘longe 
petiolatis.’ South Indian plants have pinnae 6-12 cm long, more gradually narrowed at their 
apices and either all are adnate to the axis, or at most the lcwer 2-3 pairs are free but sessile. 
The common Ceylon form therefore is distinctive in appearance and matches specimens 
from Java, which also have shortly petiolate lower pinnae, more satisfactorily than those 
from southern India. Not all specimens, however, agree in being small with short pinnae; 
Sledge 1305 resembles the Indian form in possessing pinnae 6-10 cm long and some other 
Ceylon specimens in herbaria (e.g. Schmid 1397 in BM) show an overlap in size and frond 
outline, though such fronds still differ from south Indian ones in having numerous pairs of 
free, shortly stalked pinnae. Specimens from Sikkim and Assam do not appear to be 
significantly different from those from south India. 

Shortly before Ching described Leptogramma petiolata, Iwatsuki (in Acta phytotax. 
geobot. Kyoto 19: 112-126 (1963)) published a revision of Stegnogramma, wherein 
Leptogramma is treated as a section. A consideration of the range of variation of the 
predominantly African S. pozoi and the Asiatic Leptogramma mollissima (Kunze) Ching 
led Iwatsuki to reject specific separation, but he maintained them as geographical subspecies 
and commented upon the distinctive appearance of Ceylon plants, which he considered to be 
intermediate between the two. 


50 W. A. SLEDGE 


Gymnogramma totta var. mollissima Kunze with which Ching had earlier identified 
Ceylon plants (as Leptogramma mollissima (Kunze) Ching), was described from the Nilgiri 
hills. The specimens cited by Kunze are presumed to be at Berlin but Meyer informs me that 
he has been unable to find them. It is doubtful if Iwatsuk1 is correct in identifying plants from 
China and Japan as Stegnogramma pozoi subsp. mollissima, since these have deeply lobed 
pinnae with the lowest pair often much elongated. They do not match plants 
from northern India or Ceylon and may well represent another taxon. A satisfactory 
treatment however of S. pozoi sensu lato cannot be based on herbarium material alone, since 
boundaries between described taxa are too ill-defined. This applies to African as well as to 
Asiatic plants; some specimens from eastern and southern Africa are nearer to Indian than to 
West African plants, or to specimens from Spain whence Hemionitis pozoi Lagasca was first 
described. Until experimental work affords a firmer foundation for the systematic treatment 
of the group, I believe varietal names should preferably be substituted for those of species. 
The taxon first described from the Nilgiri Hills then becomes Stegnogramma pozoi var. 
mollissima (Kunze) Sledge, comb. nov. (Gymnogramma totta var. mollissima Kunze in 
Linnaea 24: 249 (1851)). Cytological information is at present limited to chromosome 
counts on plants from Madeira and Ceylon which have shown that the former is tetraploid 
and the latter diploid. 

Both Ching (/oc. cit.) and Pichi-Sermolli (Webbia, 31:445 (1977)) maintain Lepto- 
gramma, Stegnogramma and Dictyocline. Iwatsuki (loc. cit.) and Holttum (Blumea 19 : 38 
(1971)) combine them as a single genus, Stegnogramma. The principal distinction between 
Leptogramma and Stegnogramma lies in their free and anastomosing veins respectively, but 
this difference is regarded here as inconstant. In specimens of the Ceylon species the basal 
acroscopic and basiscopic veins of adjacent lobes arch outwards and either reach the sinus 
separately or become contiguous, though not confluent, at its base. Sometimes the veins 
meet below the sinus and then they either remain contiguous but separate from one another, 
or they may fuse. Such fusions, however, if present at all, are normally so infrequent that the 
fronds are properly described as free-veined. Yet, occasionally in Ceylon plants, casual 
and irregular vein fusions may be rather numerous. A specimen of Gardner 1292 in Herb. 
Hooker at K shows such frequent anastomoses that it has been misidentified both as S. 
aspidioides Blume and S. asplenioides J.Sm. ex Ching. In one of my own gatherings (Sledge 
1340) fusions are also rather plentiful. Ching’s Leptogramma himalaica, which Iwatsuki has 
upheld, was evidently based on another such deviation from type, for, apart from occasional 
vein fusions, all the other characters attributed to it fall within the range of variation of var. 
mollissima. Morton (Am. Fern. J. 56 : 179 (1966)) reduced L. himalaica, to a subspecies of 
Thelypteris pozoi, but I believe that it is only a form. In all such plants the fusion lack the 
regularity which characterises the sections Stegnogramma and Haplogramma. A parallel, 
but reversed example, of a free-veined species within the subgoniopteroid-veined section 
Haplogramma, is provided by S. leptogrammoides K. Iwats. wherein most veins are free but 
casual and irregular fusions are frequent. Such examples vindicate Iwatsuki’s rejection of 
free, versus fusing, veins as a basis for generic separation within the group. 

There is a sheet of Stegnogramma asplenioides J.Sm. ex Ching in Herb. Hooker at K 
purporting to have come from Ceylon annotated ‘H.f. & T.’ and originally identified by 
Hooker and published (in Sp. Fil. 5 : 150 (1864)) as S. aspidioides Blume. The specimens, two 
tronds, are inseparable trom another Hooker fi/ius & Thomson sheet from the Khasia Hills, 
and since Hooker & Thomson did not collect in Ceylon and no subsequent collector has 
found S. asplenioides there, it is more likely that both sheets came from Assam. This sheet is 
the source of the erroneous records for S. aspidioides from Ceylon in the Species Filicum 
and Synopsis Filicum. Thwaites, Trimen, Beddome, Willis, Ferguson, and Wall, all either 
dismissed it as an error or placed the record in brackets in their works as being apocryphal. 
When Ching again cited it from Ceylon he was evidently copying from Hooker, and 
Iwatsuki, who has also wrongly included Ceylon in the distribution of S. aspidioides, 
appears to have copied from Ching. 


| 


Index 


Extant names are given in roman and synonyms in italic; new names are in bold, as are principal 


references. An asterisk (*) denotes a figure. 


Abacopteris thwaitesii (Hook.) Ching 47 
A. triphylla (Swartz) Ching 47 
Amauropelta Kunze 9, 10, 11 

A. bergiana (Schlechtendal) Holttum 17 
A. hakgalensis Holttum 2, 3, 4,9, 10* 
Ampelopteris Kunze 13 

A. elegans Kunze 14 

A. prolifera (Retz.) Copel. 2,4, 13,14 
Amphineuron Holttum 26, 27 

A. opulentum (Kaulf.) Holttum 2, 5, 26,27, 39 
A. terminans (Hook.) Holttum 2, 5,27, 28 
Aspidium abortivum Blume 42 

. abruptum Blume 42, 45 

. abruptum sensu Thwaites 45 
.amboinense sensu Blume 35 

_ arbusculum Willd. 43 

_articulatum (Houlst. & Moore) Lowe 45 
calcaratum Blume 14, 21, 23 

var. B. Thwaites 25 

canum Wall. 15, 17 

ciliatum Wall. 15, 17 

ciliatum Wall. ex Benth. 20 
cucullatum Blume 44 

eusorum Thwaites 42 

extensum Blume 27, 39 

. extensum sensu Thwaites 37, 39 
._flaccidum Blume 8 

. goggilodus Schkuhr 13 

. gracilescens sensu Thwaites | 1 
_hispidulum Decaisne 33 

_jaculosum Christ 32, 35 

. latipinna (Benth.) Hance 35 

. molle Swartz 32, 33 

. molle sensu Thwaites 30 

var. /atipinna Benth. 35 

obtusatum Swartz 13 

ochthodes sensu Thwaites 26, 27, 39 
opulentum Kaulf. 27 

procurrens Mettentus 3 1 

. propinquum (R. Br.) Thwaites 13 

. subpubescens Blume 32, 35 

. tenericaule (Hook.) Thwaites 7, 9 

. thelypteris 

var. squamigerum Schlechtendal 12 

. tylodes Kunze 25 

.uliginosum Kunze 7 

. unitum (L.) Swartz 44 

. xylodes Kunze 25 


MT ia A RR RR Rk 


Cheilanthes setigera Blume 7 

Christella H. Lév. 2,5, 28, 29, 31, 36, 40 

C. cylindrothrix (van Rosenb.) Holttum 32 

C. dentata (Forssk.) Brownsey & Jermy 2, 29, 
30*, 31,32, 34, 35, 36, 37, 38,40, 41, 44 

C. dentata x parasitica 36, 40 


£1 


C. hilsenbergii (C. Presl) Holttum 33 

C. hispidula (Decaisne) Holttum 2, 5, 29, 30* 
31,33, 41, 44 

C. hispidula x parasitica 46 

C. latipinna (Benth.) H. Lév. 35, 36 

C. malabariensis Fée) Holttum 29, 32, 36, 40 

C. meeboldii (Rosenst.) Holttum 1, 2, 29, 30*, 
36, 41 

C. meeboldii x parasitica 40 

C. papilio (Hope) Holttum 2, 9, 29, 37, 40 

var. repens Sledge |, 29, 37 

C.sparasitica (h.)ibevs 275,15, 29,'30*; 32533; 
34, 40, 41 

C. subpubescens (Blume) Holttum 1, 2, 29, 30*, 
31, 33, 35, 36, 37, 39, 40 

C. taprobanica (Panigrahi) Holttum 29, 39 

C. zeylanica (Fée) Holttum 1, 3, 29, 36, 39 

Cyclosorus Link |, 12 

C. arbuscula (Willd.) Ching 43 

C. contiguus (Rosenst.) Ching 33 

C. dentatus (Forssk.) Ching 32 

C. didymosorus (Parish ex Bedd.) Nayar & Kaur 
31 

C. extensus (Blume) Ching 27 

C. gongylodes (Schkuhr) Link 13 

C. hispidulus (Decaisne) Ching 33 

C. interruptus (Willd.) H. It6 2, 5,13 

C. interruptus sensu Ching 28 

C. jaculosus sensu Panigrahi & Manton 32 

C. laete-strigosus (Clarke) Ching 45 

C. latipinna (Benth.) Tardieu 35, 36 

C. meeboldii (Rosenst.) Ching 36 

C. papilio (Hope) Ching 37 

C. parasiticus (L.) Farwell 30 

C. procurrens (Mettenius) Copel. 31 

C. quadrangularis (Fée) Tardieu 33 

C. repandulus (van Rosenb.) Ching 33 

C. subpubescens (Blume) Ching 35 

C. subpubescens sensu Holttum 32 

C. sumatranus (van Rosenb.) Ching 35 

C. truncatus (Poiret) Farwell 42 

C. unitus (L.) Ching 44 


Dictyocline Moore 50 
Dryopteris Adans. | 
D. arbuscula (Willd.) Kuntze 43 
D. beddomei (Baker) Kuntze | 1 
D. brunnea Christensen 6 
D. calcarata (Blume) Kuntze 21 
var. moonii Trimen 25 
D. ciliata (Wall. ex. Benth.) Christensen 20 
D. contigua Rosenst. 33 
D. dentata (Forsskal) Christensen 32 
D. didymosora (Parish ex Bedd.) Christensen 31 
D. extensa (Blume) Kuntze 27 


32 


D. flaccida (Blume) Kuntze 8 
D. hispidula (Decaisne) Kuntze 33 
D. indica van Rosenb. 45 
D. interrupta sensu Ching 28 
D. latipinna (Benth.) Kuntze 35 
D. mauritiana 
var. gardineri Christensen 44 
D. meeboldii Rosenstock 36 
D. megaphylla (Mettenius) Christensen 45 
D. mollis (Swartz) Hieron. 32 
D. papilio (Hope) Christensen 37 
D. parasitica (L.) Kuntze 30 
D. procurrens (Mettenius) Kuntze 31 
D. prolifera (Retz.) Christensen 14 
D. pseudoamboinensis Rosenstock 35 
D. pseudocalcarata Christensen 20 
D. pteroides sensu Christensen 28 
D. quadrangularis (Fée) Alston 33 
D. repandula van Rosenb. 33 
D. setigera sensu Christensen 7 
D. subpubescens (Blume) Christensen 35 
var. major (Bedd.) Christensen 35 
D. sumatrana van Rosenb. 35 
D. tylodes (Kunze) Christensen 26 
D. uliginosa (Kunze) Christensen 7 
D. unita (L.) Kuntze 44 
D. xylodes (Kunze) Christensen 26 


Goniopteris prolifera (Retz.) C. Presl 14 
G. urophylla sensu Bedd. 46 
Grammitis totta sensu Thwaites 49 
Gymnogramma totta 

var. mollissima Kunze 50 


Hemionitis pozoi Lag. 50 
H. prolifera Retz. 14 


Lastrea Bory | 

L. beddomei (Baker) Bedd. 11, 12 
L. bergiana sensu Bedd. 15, 18 

L. calcarata (Blume) Moore 14, 21 
L. calcarata Bedd. 2, 24 

var. ciliata Bedd. 15, 17,20 
var. moonii Trimen 25 

cana J. Sm. 15,17, 18 

ciliata Hook. 20 

Jairbankii Bedd. 12 

falciloba Bedd. 15, 17, 20 
flaccida (Blume) Moore 8 
gracilescens sensu Bedd. 11 
microstegia (Hook.) Bedd. 6 
ochthodes 

var. tylodes (Kunze) Bedd. 25 
pyrrhorhachis (Kunze) Copel. 5 
sericea sensu Bedd. 15, 18, 20 

. Setigera sensu Bedd. 7 

. tenericaulis (Hook.) Moore 7 

. thelypteris 


tees OMe Wag lee hie deslicS rics ile c= 


var. squamigera (Schlechtendal) Bedd. 12 


W. A. SLEDGE 


L. tylodes (Kunze) Moore 25 

L. xylodes (Kunze) Moore 25 
Leptogramma J. Sm. 50 

L. himalaica Ching 50 

L. mollissima (Kunze) Ching 49, 50 
L. petiolata Ching 49 

L. totta sensu Bedd. 49 


Macrothelypteris (H. It6) Ching 6 
M. pyrrhorhachis (Kunze) Pic. Ser. 5 
M. torresiana (Gaudich.) Ching 2, 4, 7,9 
Meniscium parishii Bedd. 47 
M. proliferum (Retz.) Swartz 14 
M. thwaitesii Hook. 47 
M. triphyllum (Swartz) Holttum 47 
Metathelypteris (H. It6) Ching 8 
M. flaccida (Blume) Ching 2, 4, 8, 9 
var. repens Sledge 1,8 


Nephrodium Michx. | 

N. abruptum (Blume) J. Sm. 42 

N. abruptum sensu Hook. 45 

N. amboinense sensu Hook. 35, 39 

N. amboinense sensu Bedd. 36, 39, 40 
var. minor (Bedd.) Bedd. 36, 39 

N. angustifolium C. Presl 33 

N. arbuscula (Willd.) Desvaux 43 

N. articulatum Houlst. & Moore 45 

N. beddomei Baker 11 

N. calcaratum (Blume) Hook. 21 

N. ciliatum C. B. Clarke 15, 18, 20 

N. cucullatum (Blume) Bedd. 44 

N. didymosorum Parish ex Bedd. 31 

N. eusorum (Thwaites) Bedd. 42 

N. extensum (Blume) Moore 26, 27 
var. minor Bedd. 35, 39 

N. falcilobum Hook. 15 
var. B Hook. 24 

N. flaccidum (Blume) Hook. 8 

N. glandulosum 
var. /aete-strigosum C. B. Clarke 45 

N. gracilescens (Blume) Hook. 11 

N. hilsenbergii C. Pres] 33 

N. hispidulum (Decaisne) Baker 33 

N. latipinna (Benth.) Hook. 35 

N. malabariense Fée 32, 33, 36, 40, 41 

N. microstegium Hook. 6 

N. molle (Swartz) R. Br. 32 

N. molle Hook. 30, 31, 32, 41, 44 

N. molle sensu Bedd. 2, 30, 33, 40 
var. amboinense sensu Bedd. 35, 39, 40 


var. didymosorum (Parish ex Bedd.) Bedd. 31 


var. major Bedd. 32, 35, 37, 40 
N. moulmeinense Bedd. 48 
N. papilio Hope 32, 37, 38 
N. parasiticum (L.) Desvaux 29 
N. pennigerum Baker 45 


N. procurrens (Mettenius) Baker 31, 32, 40, 41 


N. prolixum (Willd.) Desv. 
var. tylodes (Kunze) Baker 26 
N. propinquum R. Br. 13 
N. pteroides sensu J. Sm. 28 
N. punctatum Parish ex Bedd. 27 
N. quadrangulare Fée 33 
N. setigerum sensu Hook. & Bak. 7 
N. smithianum C. Pres| 33 
N. tectum Bedd. 31, 33 
N. tenericaule (Hook.) Hook. 7 
N. terminans Hook. 28 
N. truncatum sensu Bedd. 42 
N. unitum sensu Bedd. 13 
N. urophyllum sensu Bedd. 46 
N. zeylanicum Fée 39 


Parathelypteris (H. It6) Ching 11 

P. beddomei (Baker) Ching 2, 4, 11 
Phegopteris distans (D. Don) Mettenius 5 
Pneumatopteris Nakai 2, 41 

P. truncata (Poiret) Holttum 5, 42 
Polypodium bergianum Schlecht. 15 
P. brunneum Wall. 6 

P. dentatum Forsskal 32 

P. distans D. Don 5, 9, 38 

P. granulosum sensu Thwaites 46 

P. late-repens Trotter 9, 38 

P. molle Jacq. 32 

P. nymphale G. Forster 32 

P. paludosum sensu Bedd. 6 

P. parasiticum L. 29 

P. pteroides Retz. 28 

P. pyrrhorhachis Kunze 5 

P. subtruncatum Bory 44 

P. tenericaule Hook. 7 

P. truncatum Poiret 42 

P. unitum L. 13, 44 

Polystichum torresianum Gaudich. 7 
Pronephrium Pres| 45, 47 


P. articulatum (Houlst. & Moore) Holttum 5, 


| 42,45 

P. gardneri Holttum 3, 5, 46 

P. nudatum (Roxb.) Holttum 48 

P. thwaitesii (Hook.) Holttum 2, 4, 47 


Pseudocyclosorus Ching 25, 41 

P. caudipinnus (Ching) Ching 15 

P. ciliatus (Wall. ex Benth.) Ching 15, 20 
P. falcilobus (Hook.) Ching 17 

_ P. ochthodes (Kunze) Holttum 26 

P. tylodes (Kunze) Ching 2, 4, 25, 26 
Pseudophegopteris Ching 5 


Pieris confluens Thunb. 12 
P. interrupta Willd. 13,28 


Sphaerostephanos J. Sm. 43, 44 


P. triphyllum (Swartz) Holttum 2, 4, 47, 48 


P. pyrrhorhachis (Kunze) Ching 2, 4,5, 9, 38 


INDEX 53 


S. arbuscula (Willd.) Holttum 2, 3, 5, 39, 43 
S. subtruncatus (Bory) Holttum 2, 5, 44 
S. unitus (L.) Holttum 2, 5, 43, 44 
Stegnogramma Blume 48, 49, 50 
S. aspidioides Blume 50 
S. asplenioides J. Sm. ex Ching 50 
S. leptogrammoides K. Iwats. 50 
S. pozoi (Lag.) K. Iwats. 49 
subsp. mollissima (Kunze) Iwats. 50 
var. mollissima (Kunze) Sledge 1,50 
var. petiolata (Ching) Sledge 1, 2, 3, 4, 49 


Thelypteris Schmidel 1, 10, 12, 24 

T. arbuscula (Willd.) K. Iwats. 43 

T. beddomei (Baker) Ching 11 

T. blumei Panigrahi 35 

T. brunnea Ching 6 

T. calcarata (Blume) Ching 21 

T. cana Ching 15, 17 

T. caudipinna Ching 15, 17, 18 

T. ciliata (Wall. ex Benth.) Ching 15, 20, 21 

T. confluens (Thunb.) Morton 1,2, 4,12 

T. contigua (Rosenst.) Reed 33 

T. dentata (Forssk.) E. St. John 32 

T. extensa (Blume) Morton 27 

T. falciloba (Hook.) Ching 17 

T. flaccida (Blume) Ching 8 

T. hilsenbergii (C. Presl) Panigrahi 33 

T. hispidula (Decaisne) C. F. Reed 33 

T. latipinna (Bentham) K. Iwats. 35 

T. malabariensis (Fée) Panigrahi 32 

T. meeboldii (Rosenst.) C. F. Reed 36 

T. paludosa sensu K. Iwats. 6 

T. palustris 
var. squamigera (Schlechtendal) Weatherby 
2 

T. papilio (Hope) K. Iwats. 37 

T. parasitica (L.) Tardieu 30 

T. pozoi (Lag.) Morton 50 

T. procurrens (Mettenius) Reed 31 

T. pseudoamboinensis (Rosenst.) Panigrahi 


315) 

T. quadrangularis (Fée) Schelpe 33 

T. repandula (van Rosenb.) Reed 33 

T. srilankensis Panigrahi 39 

T. squamigera (Schlechtendal) Ching 12 

T. subpubescens (Blume) K. Iwats. 35 

T. sumatrana (van Rosenb.) Tagawa & K. Iwats. 
35 

T. taprobanica Panigrahi 32, 33, 37, 38 

T. torresiana (Gaudich.) Alston 7 

T. tylodes (Kunze) Ching 26 

T. uliginosa (Kunze) Ching 7 

T. xylodes (Kunze) Ching 26 

T. zeylanica Ching 24 

Trigonospora Holttum 2, 4, 14, 17, 19, 20, 21, 
24,25 


54 W. A. SLEDGE 


T. angustifrons Sledge 1, 2,3, 15, 22*, 23, 24 var. angustiloba Holttum 15, 18 
T. calcarata (Blume) Holttum. 2, 3, 15, 20, T. glandulosa Sledge 1, 2, 3, 13, 15, 22* 23 
PA 22*, 23,204 T. obtusiloba Sledge 1, 2, 3, 15, 17,18, 19*, 20 
T. caudipinna (Ching) Sledge 1, 2, 15, 16*, 17, T. zeylanica (Ching) Sledge I, 2, 3, 13, 15, 
ee ZOR2 123 16* 


T. ciliata (Wall. ex Benth.) Holttum 15, 16*, 17, 
20, 21, 26 Viscum L. 49 


¢ British Museum (Natural History) 
Be: 1881-1981 


Centenary Publications 


Alfred Waterhouse 
and the Natural History 
a Museum 


% Mark Girouard 


a 6) _ During his long life (1660-1753) Sir Hans Sloane amassed a vast collection of 

____ books, manuscripts and natural history specimens. These were the foundation of 
_ the British Museum which was established in the year of his death. By 1881 the 
natural history collections had grown to such an extent that they had to be moved 
from Bloomsbury, to a new, purpose-built museum in South Kensington. There, 

_ in some five acres of land, Alfred Waterhouse had designed what is still one of 

_ London’s most outstanding pieces of architecture—The British Museum (Natural 
* History), popularly known as the Natural History Museum. 

In this book Mark Girouard, traces the chequered development of the design, 

_ describes the influences of the key personalities involved and highlights some of 
the buildings most interesting features. 


Written by one of Britain’s leading architectural historians the lively text, supported 
_ by an attractive design and lavish illustration, will appeal to all those interested in 
3 architecture and natural history. 


218 x 230 mm, 64 pages, 8 pages of full colour, illustrations, many black and white 
__ and 2 colour line illustrations. £1-75 (paper); £4-95 (hardback). 
BY Co-published with Yale University Press. 


Titles to be published in Volume 8 


The Thelypteridaceae of Ceylon. By W. A. Sledge 


Studies in the genus Hypericum L. (Guttiferae) 
2. Characters of the genus. By N. K. B. Robson 


A revision of the lichen family Thelotremataceae in Sri i 
By Mason E. Hale, Jr. 


Vascular plant collections from the Tristan da Cunha group of 
islands. By Eric W. Groves 


Printed by Henry Ling Ltd, Dorchester 


Bulletin of the 
British Museum (Natural History) 


- 
yy ign? 
a 


+ A 


Botany series Vol 8 No2 26 March 1981 


The Bulletin of the British Museum (Natural History), anatte in 1949, is issued in four 
scientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology, and 
an Historical series. 


Papers in the Bulletin are primarily the results of research carried out on the unique and 

ever-growing collections of the Museum, both by the scientific staff of the Museum and by & 
specialists from elsewhere who make use of the Museum’s resources. Many of the papers are _ 
works of reference that will remain indispensable for years to come. v1 “a 


Parts are published at irregular intervals as they become ready, each is complete in itself, ; “ 
available separately, and individually priced. Volumes contain about 300 pages and several 4 
volumes may appear within a calendar year. Subscriptions may be placed for one or more of 
the series on either an Annual or Per Volume basis. Prices vary according to the contents of _ 

the individual parts. Orders and enquiries should be sent to: ; if i 


L 
Va 
Publications Sales, - 
British Museum (Natural History), 
Cromwell Road, 
London SW7 5BD, 
England. 


aed: 


i 


World List abbreviation: Bull. Br. Mus. nat. Hist. (Bot.) 


© Trustees of the British Museum (Natural History), 1981 


ISSN 0068-2292 Botany series 7 
: Vol 8 No 2 pp 55-226 wT 
British Museum (Natural History) @ my 9 Y 
Cromwell Road 
London SW7 5BD Issued 26 March 1981 


Studies in the genus Hypericum L. (Guttiferae) 
2. Characters of the genus 


N.K.B. Robson \- 


Department of Botany, British Museum (Natural History), Cromwell Road, London 
SW7 5BD 


Contents 


Synopsis. 


\" 


Introduction . 
Relationships of Hypericum—further considerations 


(a) Delimitation of the genus 
(b) Definition of the Hypericeae . 
(c) The affinities of the Hypericeae 


2. Methodology . 


Evolution and classification 

Trends and taxonomy , 

Constructing phylogenetic diagrams 

Testingthe diagram... F 
Geographical aspects of evolutionary diagrams : 


3. Morphology . 


Habit . 
Stems . 


(a) Stem-lines and phyllotaxis 
(b) Woody stems , 
(c) Stem glands 


Vestiture 
Leaves. 


(a) Insertion and margin 
(b) Venation . 
(c) Glandularity 


Inflorescence 


(a) Form . 
(b) Bracts and bracteoles 


Sepals . 


(a) Insertion and number 


(b) Relative size, shape, union and persistence : 


(c) Vestiture 

(d) Venation . 

(e) Margin 

(f) Glandularity 

(i) Laminar 

(11) Submarginal 
(iii) Inframarginal 
(iv) Marginal 


Petals . 


(a) Insertion and number 
(b) Form . 
(c) Colour 
(d) Venation . 
(e) Glandularity 
(1) Laminar glands 
(11) Marginal glands . 


a 
| Bull. Br. Mus. nat. Hist. (Bot.) 8 (2):55-226 


55 


Issued 26 March 1981 


56 N. K. B. ROBSON 


(f) Corolla shape 
(g) Corolla movement . 
(h) Corolla (and androecium) persistence . 
Stamens 
(a) Number and arrangement 
(b) Modification of fascicles. 
(c) Androecial trends 
(d) Form and glandularity 
(e) Colour 
(f) Length and movement 
(g) Staminodial members 
Ovary . 
(a) Gynoecial elements . g F 
(b) Relationships with other whorls : 
(c) Shape and proportions 
(d) Union of styles . 
(e) Stigmas 
(f) Glandularity 
(g) Placentation 
(h) Ovules 
Fruit 
(a) General 
(b) Styles . ; 
(c) Vittae and vesicles E 
Seeds : 
(a) Number 5 
(b) Shape, size and colour 
(c) Appendages 
(d) Testa sculpturing 
(e) Embryo : 
4. Pollen morphology (By G. Cc S Clarke) : 
Introduction 
General description . ; 
Variations in pollen morphology : 
Discreteness of the pollen types F 
Distribution of the pollen types amongst the sections. 
Relationships between the pollen types. 
Taxonomic implications of the pollen morphology 
5. Biology of flower and fruit . 
Pollination . 
(a) Unspecialised flower 
(b) Insect visitors . 
(c) Floral specialisation 
Breeding systems 
Seed dispersal 
(a) Gravity 
(b) Wind . 
(c) Water. : 
(d) Animals (external) . 
(1) Birds 
(ii) Mammals 
(iii) Insects . 2 
(e) Animals (internal) . 
6. Development. 
Embryology 
(a) Normal form 
(b) Abnormalities . 
Germination 
(a) The process 


96 

97 

oF 

98 

98 

99 
100 
102 
102 
102 
102 
103 
103 
103 
105 
105 
105 
105 
106 
108 
109 
109 
109 
109 
110 
110 
110 
111 
112 
112 
ts 
115 
115 
116 
Vy 
Ty 
117 
118 
119 
119 
119 
120 
122 
123 
124 
124 
124 
124 
124 
124 
124 
125 
125 
LZ 
125 
125 
126 
126 
126 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 


(b) Factors influencing germination . 
Vegetative development : 
Floral development . 
7. Floral vasculature. 
Introduction 
Materials and methods 
Hypericum—the torus 
(a) Basic plan . 
(b) Isomerous pentamerous flowers 
(c) Flowers with pentamerous androecium and trimerous gynoecium . 
(d) Flowers with trimerous androecium and gynoecium. 
(e) Sect. Humifusoideum 
(f) Polyandrous species . 
(g) The ‘Elodes’ group . 
Hypericum—the ovary 
Other genera 
Discussion . 
8. Anatomy and phytochemistry—taxonomic implications 
Vegetative anatomy . 
Secretory system 
(a) Introduction 
(b) Distribution of secretory cavities and canals . 
(c) Composition of contents of secretory cavities and canals . 
(d) Distribution of hypericin and pseudo-hypericin. 
(e) Chemistry of hypericin and pseudo-hypericin 
Chemotaxonomy of the Hypericoideae. : 
9. Cytology and genetics . : 
Chromosomes of Hypericum . 
(a)Chromosome numbers . 
(b) Chromosome morphology 
(c) Chromosome irregularities 
Hybridisation in Hypericum . 
(a) Natural hybrids 
(b) Artificial hybrids 
(i) Sect. Ascyreia 
(a) H. x moseranum Luquet ex André (1 889) . 
(8) H. x ‘Hidcote’ , p 
(y) Other hybrids 
(11) Sect. Androsaemum . 
(111) Sect. Hypericum and intersectional hybrids . : 
(a) Hybrids not involving H. perforatum . 
(B) H. perforatum and its hybrids 
(y) Other artificial TES 
10. Distribution . 
Introduction 
Distribution of the subfamilies of the Guttiferae . 
(a) Moronobeoideae and Bonnetioideae 
(b) Hypericoideae . 
Distribution of Hypericum 
(a) Sect. 1. Campylosporus and its immediate derivatives 
(b) Sect. 3. Ascyreia and its derivatives. I. Sects 3-8 
(c) Sect. 3. Ascyreia and its derivatives. II. Sect. 9 
(d) Sect. 3. Ascyreia and its derivatives. III. Sects 10-16. 
(e) Sect. 3. Ascyreia and its derivatives. IV. Sects 17-19. 
(f) Sect. 20. Myriandra . 
(g) The Macaronesian and north- east ‘African links with the 
Mediterranean (Sects 2,21-25) . : 
(h) Old-World transcontinental distributions. I Sect. 26. 
Humifusoideum : . : , d 


58 N. K. B. ROBSON 


(i) Old-World transcontinental distributions. I]. Sects27-28. . . . . . 201 
(j) Sect: 29:,Braths=>.. 5 ee ee ee ae 201 
(k) Sect. 30. Spachium. bo wt tk ew Ss 0 206 
Interpretation of Hypericum distribution ao bea = he NS. ee Tuk: 5 208 
(a)Introduction . i ae 208 

(b) Sect. 1. Campylosporus and the Macaronesian- Mediterranean 
area . ook oe De ew as a ae 208 
(c) Sect. 20. Myriandra. Shek ee eee er 209 
(d) Sect. 3. Ascyreia and its derivatives : LO) 20 eae 209 
(e) Sects 1. Campylosporus and 26. Humifusoideum £ PFW Lineg “ae ere 212 
(f)Sect.29. Brathys . . Lon zai? (ie aie 213 
(g) Sect. 30. page : : : : : : f : ; : } : 4 214 
11. Summary. : : aie k a. ; ae : : Veneta 216 
ee Acknowledgements by fy (arnt see a Oo en ee ee 218 
13 References: wal orcmeerS ems fw @ oS, oS ec ee 0 ee 218 

Synopsis 


After a review of the relationships of Hypericum and a discussion of research methods and evolutionary 
diagrams employed, a detailed account is given of morphological variation in Hypericum, from habit to 
seed. This is followed by a summary by Dr G. C. S. Clarke of the variation in pollen structure, which is 
treated in detail elsewhere, and a discussion of flower and fruit biology (pollination, breeding systems, 
seed dispersal) in the genus. An account of development from embryo to flower is then given. A detailed 
discussion of floral vasculature in Hypericum shows how variation in this character parallels and helps 
to explain morphological variation in the flower. 

An outline of the distribution (systematic and anatomical) of the secretory system and its contents in 
Hypericum and the Hypericoideae indicates that the occurrence of hypericin and pseudo-hypericin has 
a limited but important taxonomic value. An account of variation in Hypericum chromosomes, 
including what is believed to be a complete list of known chromosome numbers in the genus, is 
followed by a consideration of hybridization in the genus. Finally, the distribution of Hypericum and 
allied genera is discussed and interpreted with recourse to the theory of plate tectonics. 

One species originally described in Hypericum (H. steyermarkii Standley) is transferred to 
Santomasia gen. noy., on account mainly of the presence in its flowers of antisepalous fasciclodes 
(sterile stamen-fascicles). One species (H. concinnum Bentham) is transferred from sect. Hypericum to 
sect. Concinna sect. noy. The pentaploid form of the hybrid H. maculatum Crantz x perforatum L. is 
designated as a nothomorph, H. x desetangsii Lamotte nm. perforatiforme (Frohl.) N. Robson comb. et 
stat. noy.; and the variegated form of H. x moseranum Luquet ex André is also given the rank of notho- 
morph, as nm. tricolor (Maumené) N. Robson stat. nov. 


1. Introduction 


The genus Hypericum having been circumscribed and subdivided in the first paper of this 
series (Robson, 1977a), this contribution is concerned mainly with the characters of the 
whole genus and its overall variation and distribution. Further papers will treat each section 
individually, and a final one will, it is hoped, provide an extended summary and include any 
necessary amendments. First, however, it is necessary to discuss the status of the anomalous 
H. steyermarkii Standley (cf. Robson, 1977a : 316) and how it affects the classification of the 
Guttiferae. 


Relationships of Hypericum—further considerations 
(a) Delimitation of the genus 


When Standley (1940) described Hypericum steyermarkii, based on a collection from 
Volcan de Santo Tomas in Guatemala (Steyermark 34760), he remarked that it had no close 
allies in either Mexico or Central America. A superficial study of the type and another 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 


SN 


vA " (VE 
VARMA _ 
(i Y HE 
‘ SS MAN Wy ip 
( 


SS 
Y MA = 


aT, 
A 


aXe 
val 
4 
ae 
Ry 
se 
i 


ve) 

Ly 

{] 

: 
ae 
yl 
Es 


x 
CX) 
as 
a 
a 
) 
z) 


=) 
ES 


Bae ? 
CAV EX ATA 
AC EDEN 
Fig. 1 Santomasia steyermarkii: (a) habit (x 0°5); (b) stem nodes and leaf scars (x 1); (c) section of 

leaf, showing venation and glands (x 2°5); (d) sepal (x 2°5); (e) petal (x 1:5); (f) flower with 

perianth removed (x 1); (g) fasciclode and adjacent stamens (x 2); (h) styles (x 2); (i) flower with 
dehiscent capsule (one petal cut) (x 1) (a-c, Matuda 2894; d-i, Steyermark 34760). 


V3 


collection (Matuda 2894, from Volcan Tacana West, Mexico) showed their undoubted 

resemblance to members of Hypericum sect. Campylosporus; but it was not until the first 
Paper was in an advanced stage of preparation that a detailed examination of these 
specimens revealed that H. steyermarkii could not be retained in Hypericum. 

The first anomalous character to be observed was the presence of vestigial staminodes 
between the stamen fascicles (Fig. If, g), and subsequently it was realised that the leaf 
venation and the occasional umbellate inflorescence were also not to be matched in 
Hypericum. In addition, H. steyermarkii has fascicles of only 11-12 stamens, petals that are 
apparently erect in flower, and often markedly auriculate leaf-bases. The last two characters 
Suggest a relationship with H. bequaertii*, the most unspecialised species in Hypericum 
(Robson, 1979); but other characters, e.g. the stamen number, do not fit into the pattern of 


*Authorities for names of Hypericum taxa will not normally be cited unless they differ from those in Robson (1977a) 
or are additional to them. 


59 


60 N. K. B. ROBSON 


variation in sect Campylosporus. In sum, therefore (1) H. steyermarkii is less specialised than 
other species of Hypericum in retaining remnants of the antisepalous androecial whorl; but 
(ii) it has certain characters that are more specialised than those of the most primitive 
Hypericum species and would therefore appear to have evolved along a different evolution- 
ary line from them. 

In Table 1 some characters of H. steyermarkii are compared with corresponding ones of H. 
bequaertii and also of H. styphelioides, the most primitive species of sect. 29. Brathys. The 
primitive condition has been marked for each character in which it has been ascertained, and 
a study of these marked characters shows that no one species is completely unspecialised 
relative to the others. 


Table 1 Characters of Hypericum steyermarkii compared with those of H. bequaertii and H. 


styphelioides (Belize form) 


H. steyermarkii 


H. bequaertii 


H. styphelioides 


Leaves elliptic (1: b=c.3) + narrowly elliptic narrowly oblong 
to narrowly oblong (1:b=4-5) 
(1: b=5-8) 
Leaf-base angustate, cuneate™*, cuneate*, 
reflexed-auriculate reflexed-auriculate not reflexed- 
auriculate 
Leaf-venation pinnate (oblique), parallel*, parallel*, 
densely reticulate not reticulate* not reticulate* 
Dark glands present, absent* absent* 
on leaves marginal and laminar 
Pellucid glands punctiform linear and punctiform 
on leaves punctiform* 
Inflorescence 1(2-4)-flowered 1-flowered 1-flowered 
(umbellate if more 
than one) 
Dark glands on present, marginal absent or present, absent 
sepals and submarginal marginal and 
inframarginal 
Petals erecta. erect*, spreading, 


subsymmetrical*, 
without lateral 


asymmetrical, with 
rounded lateral 


asymmetrical, with 
acute lateral 


apiculus* apiculus apiculus 
Stamens per 11-12 ee COs 
fascicle 
Staminodes present* absent absent 
Styles free* c.3/4 coherent free* 
Seed shallowly carinate not carinate slightly carinate, 
with apical wing or winged not winged 


* = primitive state of character 


Although the inflorescence of H. steyermarkii is usually a single flower with bracteoles at 
the base of the pedicel, in the type specimen it is sometimes 2—4-flowered. In such 
inflorescences the bracteoles are crowded together, i.e. the inflorescence is umbellate, a state 
that is otherwise unknown in primitive species of Hypericum. The leaf of H. steyermarkii, 
broad but narrowing to a sessile reflexed-auriculate base, contrasts with those of species of 
Hypericum proper, where the only species in which such leaf-bases are at all strongly 
developed are the most primitive species of the most primitive section, Campylosporus (H. 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 61 


bequaertii, H. revolutum and H. lanceolatum). In these species, however, the leaf is narrow, 
and the functional secondary veins are much fewer than they are in the leaf of H. 
steyermarkii (cf. Fig. 9d, e) (p. 77). These secondary veins and the intervening glandular 
canals are almost parallel with the leaf margin, as are those towards the base of the leaves of 
two Bonnetiaceous genera, Neblinaria and Neogleasonia, recently described by Maguire 
(1972) (Fig. 9a, b). Melville (1969) described the leaf-venation of some specimens of H. 
bequaertii (subparallel veins from a pedate base) as gangamopteroid, a state which he regards 
as primitive in the Angiosperms and as having given rise to the glossopteroid (pinnate) state. 
According to his later terminology of leaf architecture (Melville, 1976), such venation is 
pedati-flabellate (cf. H. styphelioides, Fig. 9f); whereas the venation of the above-mentioned 
Bonnetiaceous genera is lirate, an early stage of the more advanced pinnate system. On the 
other hand, whereas in Hypericum the leaf always has a complete midrib, in Neblinaria, the 
midrib is well-developed only at the base, a very primitive state (Fig. 9a). In contrast to 
Melville, Hickey & Wolfe (1975) regard pinnate venation as probably primitive in the 
Dicotyledons and would describe the venation of H. bequaertii as parallelodromous (Hickey, 
1973), that of Neblinaria and Neogleasonia (Fig. 9b) being craspedodromous and thus in 
their view more primitive. 

The evolutionary trends in the leaf venation of the Bonnetiaceae-Guttiferae suggest that 
- Melville is right and that the flabellate type of venation has given rise first to the lirate type 
and then to more typical pinnate types. In this sequence the major venation of H. 
steyermarkii is at least partly pinnate, approaching that of, say, Bonnetia. More advanced 
species of Hypericum proper also have pinnate leaf-venation, but this has evolved in a 
different way (see below, p. 80). A further difference occurs in the densely reticulate tertiary 
venation in H. steyermarkii, which is an advanced character comparable with, but quite 
different in origin from, the vein-reticulations in H. roeperanum and H. gnidiifolium, 
relatively advanced members of sect. 1. Campylosporus (cf. Robson, 1979). 

The dark gland dots in the leaves and sepals of H. steyermarkii (Figs 1c, d, 9d) are com- 
parable with those of H. bequaertii (Fig. 9c), although the laminar ones recall the more 
superficial (less deeply immersed) glands of some Bonnetiaceous genera (e.g. Bonnetia). On 
the other hand, the numerous pellucid gland dots appear to be homologous with the minute 
ventral pellucid gland dots of H. bequaertii and with those of H. styphelioides (Fig. 9f), not 
with the scattered pellucid gland dots in the vein areoli of e.g. H. roeperanum. These 
originate from functionless veins, as will be shown later. 

With two isolated localities in Central America, H. steyermarkii has a quite different 
distribution from that of sect. Campylosporus, which is confined to the African mainland 
and some adjacent islands. Although sect. Brathys is Central and South American, H. 
styphelioides and its other species are even less similar to the Mexico—Guatemalan plant. 
Despite all these differences, H. steyermarkii is in general much more closely related to 
_ Hypericum proper than it is to any other genera. Its appropriate position would therefore 
seem to be as a second genus of the Hypericeae cognate with Hypericum. The new genus is 
named Santomasia after the type locality in Guatemala, Volcan Santo Tomas. 


Santomasia N. Robson, gen. nov. 


| Hyperico L. affine, a quo staminorum fasciculis sterilibus reductis 5 inter eos fertiles insertis, 
_ inflorescentia | -florata vel fasciculata, differt. 

Arbores vel frutices glabri. Folia opposita integra exstipulata penninerva glandulis 
_pellucidis nigricantibusque obsita. Inflorescentia terminalis, 1-florata vel 2-4-florata 
_ fasciculata. Sepala 5, quincuncialia glandulis pellucidis nigricantibusque obsita. Petala lutea 
_ subsymmetrica erecta post anthesin persistentia. Staminorum fasciculi 5, filamentis basi 
breve coalitis anthera glandulifera ferentibus, post anthesin persistentia. Staminodia 
(staminorum fasciculi steriles) 5, parva, inter staminorum fasciculos fertiles inserta. Ovarium 
5-loculare loculis oo-ovulatis, stylis elongatis liberis stigmatibus parvis. Fructus capsularis 


62 N. K. B. ROBSON 


valvis longitudinaliter anguste vittatis, ab apice dehiscentibus. Semina numerosa, cylindrica 
vade carinata, terminale alato-appendiculata. Grana pollinis tricolporata sphaeroidalia, 
colpis longislatis, endoaperturis lalongatis, ornamentatione microreticulati, munita. 


TYPUS GENERIS: S. steyermarkii (Standley) N. Robson 

Genus adhuc monospecificum. 

S. steyermarkii (Standley) N. Robson, comb. nov. 

Hypericum steyermarkii Standley in Publs Field Mus. nat. Hist. Chicago (Bot.) 22: 160 (1940); 
Standley & Williams in Fie/diana (Bot.) 24 (7) : 50 (1961). 


Shrub or tree 2-9 m high, much branched, with branches ascending; bark grey, fissured. 
Stem 4-lined, slightly flattened and orange-brown when young, soon 2-lined to terete with 
rather swollen nodes; internodes 4-15 mm long, much shorter than the leaves, with 
numerous small reddish or dark glands. Leaves sessile; lamina 25-70 x 9-24 mm, elliptic, 
acute or apiculate, the margin plane, the base cuneate to angustate or reflexed-auriculate, 
paler beneath, not glaucous, chartaceous, the lower ones soon deciduous, with 9-14 lateral 
veins, the lower ones arcuate, the middle ones spreading, the upper ones flexuous and 
ascending, and a uniform densely reticulate tertiary venation, the laminar glands very small, 
punctiform, pellucid and sometimes also dark, the inframarginal gland-dots pale, dense, 
alternating with more widely spaced dark dots. Inflorescence 14-flowered; pedicel 9-20 mm 
long, sometimes dark-gland-dotted; bracts foliar but smaller and broader; flowers c. 3-4 mm 
in diam., erect. Sepals free, imbricate, equal, reflexed in fruit, 4-7 x 2-5-3 mm, broadly 
ovate, acute to obtuse, with margin entire or eroded-glandular-ciliate towards the apex, with 
midrib obscure; laminar glands pellucid, numerous, linear, submarginal gland-dots dark, a 
few subapical, the rest dispersed; marginal dark glands few, present especially towards the 
apex, sessile or on cilia. Petals deep yellow, 17-21 x 9-11 mm, c. 44:5 x as long as the 
sepals, obovate, without lateral apiculus, the margin distally dark-glandular-ciliate, some- 
times with a few large submarginal dark glands, with pellucid linear glands interrupted 
distally, erect at anthesis, persistent in fruit. Stamen fascicles c. 10 mm long, c. 0°5 x as long 
as the petals, each with 11-12 stamens, persistent in fruit; filaments shortly united, relatively 
stout (cf. Hypericum); anther gland amber. Staminodial fascicles c. 1:5 mm‘ long, oblong, 
truncate, spreading, fleshy. Ovary c. 8 x 3 mm, narrowly ovoid; styles 3°54 mm long, erect 
to near apex. Capsule 12-14x 10mm, ovoid-cylindric, with valves finely longitudinally 
vittate. Seeds pale brown, cylindric, c. 2-2°5 mm long, shallowly carinate, with terminal 
winged appendage, finely shallowly linear-reticulate. 


HasiTaT. ‘In forest of pine and Abies’, 2500-3700 m. 
DISTRIBUTION. Guatemala (Quezaltenango) and Mexico (Chiapas). 


SPECIMENS. Guatemala: Quezaltenango, Volcan Santo Tomas, 2500-3700 m, 22.1.1940, Steyermark 
34760 (F, holotype), 34774 (F). Mexico: Chiapas, Volcan Tacana West, 2800 m, 30.11.1939, 
Matuda 2894 (F, MEXU, MICH, NY, US), S-228 (MICH). 


(b) Definition of the Hypericeae 


If one accepts the Englerian classification of the Hypericoideae as a subfamily of the 
Guttiferae, then the inclusion of Santomasia in the Hypericeae necessitates an enlargement 
of the circumscription of this tribe (cf. Robson, 1977 : Table 2). The characters of the three 
tribes of the subfamily as now constituted are shown in Table 2. 


(c) The affinities of the Hypericeae 


From Table 2 it is clear that the three tribes of the Hypericoideae are very different, so that 
there is no difficulty in assigning a given species to one of them. It is equally clear, when 
other subfamilies of the Guttiferae are considered, that none of them can have given rise to 
the Hypericoideae. The nearest one in many respects is the Kielmeyeroideae, which Maguire 
(1972) has (rightly in my opinion) united with the Bonnetiaceae. The family so formed has 


SPUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 


63 


Table 2 Characters of the tribes of the Hypericoideae 


Tribe 


Genera 


Leaf venation 


Inflorescence 
branching 
Sepals 


Petals 


Fasciclodes 


Stamen fascicles 


Styles 
Placentation 


Fruit 


Dehiscence 


Seed 


Cotyledons 


Vismieae 

Vismia Vand. 
Psorospermum Spach 
Harungana Lam. 


pinnate 
(co-arcuate) 


thyrsoid, 
racemiform 
5, quincuncial 


5, white or greenish 
to yellow or orange, 
nearly always 
internally villous, 


without or rarely with ligule 


5 


5, free 


5, free 
5, axile 


baccate (co—5-seeded) 
or drupaceous 

(5 pyrenes) 
indehiscent 


not winged or carinate, 
sometimes gland- 
dotted 

equal, straight 

or unequal, curved to 
inrolled 


Cratoxyleae 
Eliea Cambess. 
Cratoxylum Blume 
Thornea Breedlove 
& McClintock 
Triadenum Rafin. 
pinnate 
(co-arcuate) 


thyrsoid, 
racemiform 

5, quincuncial 

or irregularly 
imbricate 

5, white or greenish 
to pink or crimson, 
glabrous, without 
or with ligule 


3 (4-5) 


5, united2+2+1 
(very rarely 
2+14+1+41 

or free) 

3 (4-5), free 


3 (4-5), axile 
capsular 
loculical or 
septicidal 

or both 
winged or not, 


not gland-dotted 


equal, straight 


Hypericeae 
Santomasia 

N. Robson 
Hypericum L. 


flabellate (pedati- 
flabellate) to 
pinnate (co-arcuate) 
dichasial or 
thyrsoid, dichasial 
5, quincuncial 

or 4, decussate 


5-4, yellow to 
orange, often + 
suffused with red, 
glabrous, without 
or very rarely with 
ligule 

0 or very rarely 

5 or3 

5, free or united 
2+2+1 (rarely 
2+1+1+1)or54 
wholly united 
5-2, free or + 
united 

5-2, axile to 
parietal 

capsular or rarely 
baccate or 
tricoccoid 
septicidal or rarely 
indehiscent 


winged or carinate 
or not, not gland-dotted 


equal, straight 


Data from Engler (1925), Perrier de la Bathie (1951), Ewan (1962), Bamps (1966), Gogelein (1967), Baas (1970), 


Robson (1974). 


characters that bridge the gap between the Guttiferae and Hypericaceae of Bentham (1862) 
(who excluded the Kielmeyeroideae from both families), on the one hand, and the Theaceae 
sensu stricto (i.e. the subfamilies Camellioideae and Ternstroemioideae) on the other. For 
example, in it are included genera with latex (e.g. Kielmeyera, Neotatea) and without it (e.g. 
Bonnetia), with leaves alternate (e.g. Archytaea) or opposite (e.g. Marila)—even rarely 
stipulate (Mahurea), without oil-glands (e.g. Archytaea) or with them (e.g. Caraipa), and 
with petals that are imbricate (quincuncial) in bud (e.g. Neogleasonia) or contorted 


(Kielmeyera). 


64 N. K. B. ROBSON 


Although these intermediate characters would suggest that the Bonnetiaceae is primitive 
relative to the Guttiferae sensu lato, they are accompanied by more specialised characters 
which indicate that the relationship is not a simple one. Thus, the only genera in the 
Bonnetiaceae in which the antisepalous stamen whorl occurs are Ploiarium and Archytaea 
(Kobuski, 1950), where it is sterile and much reduced in size. These genera are also the only 
ones with a 5-merous gynoecium. On the other hand, their leaf-venation is, as we have seen, 
more advanced than the most primitive state in Hypericum. We can say, therefore, that the 
Bonnetiaceae and the Guttiferae arose from the same stock, but not that the former gave rise 
to the latter. Even if one goes further back along the evolutionary line of the Dilleniidae to 
the Dilleniaceae itself (pace Corner (1976), who denies that they are on the same line), it is 
not possible to find a direct ancestor of the Guttiferae. Although the Dilleniaceae includes 
species with floral characters that are more primitive than corresponding ones in the 
Guttiferae (e.g. the apocarpous ovary), yet there are other respects in which all its members 
are more advanced than the most primitive species of Guttiferae and Bonnetiaceae (e.g. the 
pinnate leaf-venation (Rury & Dickison (1977)). It would therefore seem, not surprisingly, 
that the ancestral species linking the Dilleniaceae, Bonnetiaceae and Guttiferae are extinct 
or still to be discovered. 

To return to the Hypericoideae, whereas the most primitive leaves in Vismia (e.g. in V. 
japurensis Reichardt) resemble those of the Bonnetiaceous Marila, for example, and the 
stellate indumentum present in most of the Vismieae is comparable and possibly homo- 
logous with that in Caraipa, its floral characters are partly more primitive and partly more 
advanced—more primitive in the better-developed staminodial whorl, more advanced in 
having internally villous petals, stamen filaments in each fascicle united above the middle, 
and baccate fruit. The pinnate leaf-venation in the genera of the Cratoxyleae is likewise 
advanced relative to the most primitive state in the Bonnetiaceae (‘parallel’); and the usual 
trimery in the inner floral whorls is another relatively advanced character, which is associ- 
ated with erect petals and united filaments, as in the Vismieae. Yet, here again the 
staminodial whorl is better developed than in Ploiarium and Archytaea. The Hypericeae, on 
the other hand, are at the same evolutionary stage as the Bonnetiaceae as regards the 
development of the staminodial whorl, apart from the ‘Elodes’ group, in which, as has 
already been shown (Robson 1972a, 1977a), it appears to have been re-acquired. The 
presence of five fascicles of stamens is a primitive character within the Hypericeae (Robson, 
1977a, see also below, p. 100), and such fascicles occur in the Bonnetiaceae only in 
Ploiarium and Archytaea, the androecium of the other genera being at the evolutionary level 
of, say, Hypericum sect. 20. Myriandra. But, even in these two genera, the filaments are 
united above the base. We can therefore conclude that no tribe of the Hypericoideae is 
wholly advanced in character with respect to the Bonnetiaceae, even though the latter is in 
general more primitive. 

In a recent study of the wood anatomy of the Bonnetiaceae, Baretta- Kuipers (1976) upheld 
Maguire’s view that it is a family quite distinct from the Theaceae. Excluding Kielmeyera, 
which is somewhat different from the others, she found that in wood structure the genera 
together form a gradual transition from Theaceae to Guttiferae. She also suggested that 
Poeciloneuron Beddome, a genus of two species from south-western India placed by Engler 
(1925) in the Guttiferae—Calophylloideae, might be incorporated into the Bonnetiaceae. 

It seems clear, therefore, (i) that the relationships of the Hypericoidae to the Bonnetiaceae 
are just as close as they are to other subfamilies of the Guttiferae and (i1) that the circum- 
scription of some of these subfamilies is still not wholly certain. In these circumstances I 
would suggest that the Bonnetiaceae be incorporated in the Guttiferae as a subfamily 
co-ordinate with the Hypericoideae, Calophylloideae, Moronobeoideae, Clusioideae and 
Garcinioideae (probably distinct from the Clusioideae), thus making the Guttiferae an entity 
which is natural and well separated from adjacent families. 

Kubitzki (1978), in his study of Caraipa and Mahurea, comes to similar conclusions 
regarding the inclusion of the Bonnetiaceae in the Guttiferae and further points out that the 
former differ from the Theaceae sensu stricto in having rubiaceous stomata, in certain wood 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 65 


characters (cf. Metcalfe & Chalk, 1950) and in the occurrence of xanthones (Lima, Gottlieb 
& Lins Mesquita, 1972). The last character, in particular, is typical of the Guttiferae. 


2. Methodology 


Before we turn to detailed consideration of Hypericum, a short outline of the philosophical 
basis of the work and the consequent methods employed in it may be helpful. 


Evolution and classification 


In the absence of a useful fossil record—and, with the possible exception of Pliocene seeds 
(see e.g. Reid, 1923), there are no known fossils of Hypericum—evolutionary data must be 
obtained from recent forms. Self-evident and a truism though this statement may be, there is 
no general agreement yet about the validity of data obtained from this source. One still 
encounters the affirmation that all such data are merely speculative and thus ‘unscientific.’ 
On the other hand, the publication of numerous papers in which evolutionary aspects of 
contemporary organisms are discussed indicates that many biologists do not share this view. 
Much has also been written about the relation between evolution and classification, ranging 
from a denial that they are related (’phenetic’ classifications) to the publication of 
‘evolutionary’ or ‘cladistic’ classifications. Workers who take an intermediate position in this 
debate (e.g. Mayr, 1968) have suggested that a classification should reflect the evolutionary 
history of the group as closely as possible. There is no group of recent organisms of which all 
the members still exist, and so all classifications of such organisms involve hypotheses 
concerning the relationships within the group. Incidentally, the same can be said of classifi- 
cations that do include fossils. It is the existence of gaps in the range of variation due to 
absent members of a taxonomic group that enables us to classify at all. Classification is thus 
an analytical process. In order that the analysis may result in taxa that are monophyletic and 
that together reflect the apparent course of evolution in the group, it is desirable first to 
perform a synthesis, i.e. to construct a phylogenetic scheme. ‘First look for the similarities, 
then the differences.’ By so doing, the chances of misclassification are minimised, and the 
resultant classification is more likely to be natural, i.e. to represent the actual variation 
pattern in hierarchical categories. Subsequently the gaps in variation, and hence the limits of 
the taxa, can be determined. 

From the above considerations it will be seen that the problem in Hypericum classifi- 
cation, as it would be in that of all wholly recent taxa, has been to obtain evolutionary data 
and construct a phylogenetic scheme using non-fossil data only. 


Trends and taxonomy 


One of the most serious criticisms that are levelled at herbarium taxonomy is that it is 
concerned mainly with adult structures, whereas evolution involves repeated ontogenies. 
Although the use of more-widely-based evidence goes some way towards answering this 
criticism, it is still true that classical taxonomy tends to be static in concept, whilst evolution 
is a dynamic process. For this reason I have reservations about the usefulness of neo- 
Adansonian methods of classification—and indeed about all methods involving maximum 
correlation of individual characters. Being based on the static ‘character’ concept, they 
cannot be relied upon to produce a natural classification (cf. Mayr, 1968). On the other hand, 
the dynamic aspect of evolution is reflected in character trends, i.e. directional changes in a 
character through time. Any taxon has both stable and variable characters; and whereas 
character-correlation methods tend to treat only the stable ones satisfactorily, the variable 
a are those that produce the trends and can therefore be important indicators of 
affinity. 

Trends can be observed equally well in fossil and recent material; and there seems to be no 
good reason why spatial trends (i.e. those observable in geographically displaced taxa) should 


66 N. K. B. ROBSON 


be regarded as less ‘valid’ than those derived from a chronological sequence of fossils. The 
difference between the two types is merely that the trends are manifest in a reasonably 
complete fossil sequence, whereas they have to be worked out for contemporaneous taxa. Of 
course, ‘working out’ is also necessary for all but the most complete series of fossils, so that 
the need for hypotheses is rarely less for them than it is for recent organisms. If the validity of 
interpreting character trends among contemporaneous taxa as evolutionary trends is 
admitted, there still remain the problems of recognising them and interpreting their direction 
correctly. These topics have recently been discussed in detail by Stebbins (1974; ch. 70), who 
pointed out that, although trends in single characters are sometimes reversed, if trends are 
based on simple alterations of the development pattern, the more complex and multi- 
factorial the basis of a character is, the less reversible it will be. It follows that, if one 
correlates trends rather than characters, then chances of being misled by convergence or 
parallelism are minimised and general evolutionary directions can be accurately established. 
It is usually more difficult to establish the true direction of a trend than to recognise the trend 
itself. Hennig (1966a) and Stebbins (1974) both discuss the criteria by which trend direction 
may be discovered; but during the present work the direction of a given trend had often had 
to be assumed initially. Subsequent correlation with other trends and distributional data has 
almost always indicated whether or not the original assumption was valid. By this method it 
has usually been possible to build up a coherent broad evolutionary picture, with a fair 
degree of confidence in its accuracy. 


Constructing phylogenetic diagrams 


The size of Hypericum (over 400 species) would preclude the extensive use of numerical 
methods of character assessment, even if it had been thought desirable to use them, and so 
the process of recognising and correlating trends has been largely one of inspection. If a 
search is made for the ‘nearest neighbour’ of each species, then an affinity diagram such as is 
shown in Fig. 5 (p. 70) will result. This diagram differs from a Hennigian sister-group one in 
showing some multiple affinities because it is based on overall resemblance. Hennig is no 
doubt correct in assuming that, for practical purposes, each monophyletic taxon originates at 
one time from one ancestral taxon, so that only two sister-groups are thereby formed. Never- 
theless, variation in (say) a species may occur in several morphological and geographical 
directions at once. If it is split into more than two parts and speciation subsequently occurs, 
each sibling species will include a different part of the variation of the original species. The 
actual order in which the parts are split off may be evolutionarily interesting; but to make it 
the basis of classification, as adherents of the Hennigian school advocate, seems to be 
unnecessary. 

The published sectional classification diagram of Hypericum (Robson, 1977a: figs 1, 2), 
which is reproduced in a slightly improved form in Fig. 2, shows examples of sections with 
multiple derivatives and more than one hidden example of paraphyly (i.e. of a taxon that 
does not include all the derivatives of its ancestral group). One paraphyletic line comprises 
sects 27. Adenosepalum and 28. Elodes, as the latter (monotypic) section is apparently 
related to relatively advanced members of the former. Nevertheless, the transition from open 
pollination to specialised pollination, with accompanying profound morphological modifi- 
cations (Robson, 1977a: 298-302), provides a conspicuous gap in variation which, I think, 
should be reflected in the classification. Admittedly the rank to adopt is a matter of judge- 
ment, but so it is in many taxonomic decisions. With regard to the sections with multiple 
derivatives, the detailed relationships of each one can be represented in a cladistic diagram 
such as Fig. 3. The broken line defining sect. 1. Campylosporus includes all those species 
with regularly pentamerous flowers, persistent petals and stamen fascicles, and at least partly 
united styles. All gaps in variation to the outside of this line are considerable and easily 
recognised, with the exception of that separating sect. 1. Campylosporus from sect. 3. 
Ascyreia. Here two of the key characters (1) persistent versus deciduous petals and stamens 
and (ii) united versus free styles are not completely reliable in H. socotranum. Nevertheless, 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 67 


Trees 


2 ite me igre 


i a Clee ie RY 


Fig.2 Affinities of sections of Hypericum, slightly modified from Robson (1977a: figs 1, 2). 


the complete geographical segration of the two taxa and the fact that united styles recur only 
in advanced species of sect. Ascyreia (e.g. H. monogynum) allow one to recognise them as 
sections. The forked line to sects 24 and 25 indicates that the relative advancement of these 
sections is not clear. Thus, although the vegetative parts in sect. 24. Heterophylla (H. hetero- 
phyllum) are more advanced (e.g. in having the leaf-differentiation from which it gets its 
name) and its distribution suggests a derived relative position, the flowers are homostylous, 
whereas in sect. 25. Adenotrias they are heterostylous. The evolutionary diagrams of sections 
(e.g. Fig. 5) have been produced by stages, in a way rather similar to building up a jigsaw- 
puzzle picture. Certain primitive/advanced relationships are usually obvious; but, even if the 
trend directions in them are not clear at first, small separate ‘pieces’ of the diagram can be 
constructed on the basis of overall similarity. In theory, the whole diagram could be 
produced in this way (Fig. 4); but in practice it and others like it have resulted from a con- 
sideration of trends at an early stage. Thus its conversion into an evolutionary hypothesis by 
the insertion of (i) directional indications and (ii) the ‘nearest neighbour’ or ‘sister group’ to 
the whole section (Fig. 5) was not an entirely separate process. The two stages, however, 
show some similarity to the ground plan/divergence diagrams produced by Wagner (1952a, 
b) and subsequently elaborated by him and others (e.g. Wagner, 1969; Farris, 1970). The 
first-stage (interrelationship) diagram corresponds to the ‘Wagner network’ and the second 
(cladal) to the ‘Wagner tree’ (Nelson & Van Horn, 1976), but a numerical basis is lacking (for 
the reason stated above); and in my opinion it is unnecessary. Although in the early stages 
the diagram usually consists of separate linear series of taxa, based on trends in few 
characters, as other trends are gradually correlated, these linear series become ‘fixed’ in the 
picture. One therefore has a reasonable amount of confidence that the final diagram does 


show actual relationships, at least in broad outline; and it is possible to test its accuracy to 
some extent. 


68 N. K. B. ROBSON 


sect.25 Adenotrias 


sect. 22 sect.24 Heterophylla 


Arthrophyllum 


sect.23 
sects 4-19 _ 
sect.2 Psorophytum A Triadenioides 


sect.21 Webbia sect. Ascyreia 


H. socotrantimag 
sensu lato N 


sect. 20 
Myriandra -* 


H. madagascariense 


H. synstylum H. revolutum 
; subsp. revolutum 
'‘H. roeperanum 

Soene bt Se H. quartinianum 


rae a 
. 


sect.27 Adenosepalum °, 


H. lanceolatum 


subsp. lanceolatum 


me 


H. lanceolatum 
1 


subsp. angustifolium 


7 
a 
va 
- 
va 


y * H. revolutum os 
Seeu 28 Elodes ” subsp. keniense wi 
a 7 uw 
t.1 ye 
sect.29 SEC is 
Brathys sect.30 : Campylosporus | ’ 
Spachium | 4 
1 


- 


SECC a6 


S. steyermarkii Humifusoideum 


“Et bequaertin —-— 


oa 
- - 
~ - 
~ - 
mies aie 
SS a 


[Santomasial [Hypericum] 


Fig. 3 Sect. 1. Campylosporus, showing its limits (dotted line) and the interrelationships within 
and outwith the section. For revised position of H. gnidiifolium, see footnote on p. 187. 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 69 


H. edisonianum 


H. hypericoides H. Fey H. ellipticum 
H. suffruticosum —— H. crux-andreae H.dolabriforme ra adpressum 


Mic /. 


H. kalmianum H. oS H. myrtifolium H. sphaerocarpum 
~~ apocynifolium H. pao 
H. buckley! H. prolificum H. nudiflorum H. cistifolium 
H. galioides 
H. densiflorum —H. lobocarpum eS H. fasciculatum — H. chapmanii 


H.reductum 


H. lissophloeus H. exile 


H. nitidum 


H. brachyphyllum H. |loydii H. limosum 


Fig.4 Relationships in sect. 20. Myriandra, based on overall similarity (Stage I). 


Testing the diagram 


The earlier efforts to produce a classification of Hypericum (Choisy, 1821, 1824; Spach, 


1836a, b, c; Keller, 1893, 1925) resulted in schemes that were partly gradal, i.e. based on 
corresponding stages of evolutionary development, not on descent; and the general aim of 
the present studies has been to convert these gradal schemes into a cladal one. The lines on 
the evolutionary diagrams, therefore, all indicate trends, and it should be possible to indicate 
the points of corresponding character-change (i.e. gradal points) along each line. Joining 
these points should produce concentric ‘contour’ curves or circles, depending on whether the 
primitive taxon is on the margin of the diagram or not (Robson, 1977a: Fig. 1). If, however, 
(a) the character change has taken place independently more than once (i.e. if the advanced 
character is polyphyletic), or (b) trend-reversals have occurred or (c) a trend has been 
misinterpreted, then two or more non-concentric circles will result. Thus Fig. 6 shows that in 


_ Hypericum there has been a trend from five styles to three or sometimes two; but in four 


sections (13, 15, 26, 30) there has been a reversion to five styles. Of course, as with 


_ hypotheses in general, the diagram can also be checked by adding data not used in its 


construction. 
In conclusion, it should be stated that, although the above method of constructing and 
testing evolutionary diagrams was developed independently, no part of it is wholly original. 


70 N. K. B. ROBSON 


Sect.1 Campylosporus 


H. hypericoides 
(Mass - Fla-Tex-III, 
Mex.-Hond.Rep, W. Indies) 


H. synstylum 


H. suffruticosum ae edisonianum 


i (NC- Fla- v DA (tia) 
\ 
H. kalmianum H. frondosum 3 
m= Bess ss ee rg ea eh aa 
(Que, Ont, oe Mich, (NY, eres Ge) 
Ind, NY) es Sx i ca 


H. prolificum H. oe: eae. 
(Penn =Ge-OKla= lowa - Mich)/ lor Fla-Tex- Ky (Ge, Fla,Cuba) 
re 
| 
Ei bucklewil Gu: myrtifolium H. ellipticum 
(NC, SC, Ge), ECS es (Nfld-W Va -Tenn-Wis - Ont) 
/ 


Alt 
ii | 
ratyee \ 
H. densiflorum 
(NJ-Ge-Al-Tenn) 

/ 


H. apocynifolium BAey We 
(Tex,Ark, La, Ge, Fla) 
Se 


H. dolabriforme 
(Ky, Tenn, Ge) 


He H. adpressum 

! (Mass-Ge-Tenn-Mo-Ind) 

H. sphaerocarpum 

H. Alioiagtes (Ohio-Al-Okia-Wis) 

(Nc-Fla-Tex) j | J NS 
i) 


\|\. H.nudiflorum 
/ \ 
/ 

H. lobocarpum! 


(Tenn-Miss-tIl) 


x =- 


~\<--- 


ST _o 
H. lissophloeus H. fasciculatum —— iH. chapmanii ' 
(Fla) (SC-Fla-Miss) (Fla) 
.reductum 
NC,Ge, Fla 


— Sepals persistent »deciduous 


H. exile 
(Fla) =— Petals 5—»4 


=—=—— Stamens deciduous—> 
H.nitidum persistent 
-Fla ba,Belize 
(NC , Cuba, we aN ——-—— Styles 5 fewer 
4 ucun Styles more than 2—» 2 
H.brachyphyllum H.limosum — H. lloydii 


Ge, Fla, Al, Miss) (Cuba) (Nc, Sc) —-—.- Leaves articulated at base —» 
not articulated 


+++— Inflorescence branching 
primarily basipetal—> 
acropetal 


Fig. 5 Relationships in sect. 20. Myriandra, showing evolutionary trends and also some charac- 
ter limits (Stage II). 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 71 


remainder 


Fig.6 Variation in style number in Hypericum. 


| Thus, the resemblance of the diagrams to Wagner trees has already been pointed out; and 
character-limits have been shown on evolutionary diagrams by several authors (e.g. 
_Kubitzki, 1969; Dahlgren, 1977). Nevertheless, the method has proved to be useful in 
studying the cladistic relationships of a large number of species without resorting to time- 
consuming numerical methods; and from the resultant diagrams it is usually relatively easy 
_to decide how to classify and key out the group. 


Geographical aspects of evolutionary diagrams 


Because evolution usually takes place in space as well as time, there is a distributional 
component to most evolutionary trends. It is therefore possible to retrace on a map changes 
that have taken place over a period of time. This is one way of checking an evolutionary 
_hypothesis—it must make geographical sense. Thus, in an evolutionary line one is likely to 
find that the most primitive taxon has a restricted distribution, the early derivatives are more 
widespread and frequently occur in disjunct areas, and the most recently evolved taxa either 
are the most widespread or show signs of having evolved from a widespread taxon (neo- 
-endemics). A trend in sect. 26. Humifusoideum (H. natalense—» H. wilmsii—* H. peplidi- 
folium) provides a good example (Fig. 7). 

Hennig (1966a, b) proposed three Biogeographical Rules, which have been formalised and 
extended by Ashlock (1974). These are intended to summarise the relationships between 
phylogeny and geography. The trend shown in Fig. 7 is an illustration of the Progression 
Rule: The direction of progression of a geographical sequence of taxa is indicated by a series 
of increasingly derived synapomorphic characters. If a group of organisms fits this rule, then 
it is an indication that its history has not been seriously disturbed e.g. by extensive 
hybridisation. The other two Hennigian rules, the Phylogenetic Intermediate and the 


We N. K. B. ROBSON 


Under 500 m| 
500 —1000m} 
1000—1S00 mf: 
1500—2000 m|[]if : 


Over 2000 mE=4 
5 — 


10 1S 0 U 40 45 30 


Fig. 7 Distributions of African members of sect. 26. Humifusoideum. @ H. natalense, a H. 
wilmsii, @ H. peplidifolium. 


Multiple Sister-Group Rules, which refer to relationships between taxa on different 
continents, and Ashlock’s (1974) fourth rule, the Drift Sequence Rule, which relates these to 
the theory of Continental Drift, concern topics that will be considered later (Chapter 10). 
In addition to these Biogeographic Rules, the following three more or less self-evident 
points have been borne in mind: 
(i) The more primitive the taxon, the more likely it is to have had a past distribution 
markedly different from its present one in location and/or extent. 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 73 


(ii) Anomalies in distribution (such as wide disjuncts) that cannot be due to climatic or 
intracontinental physiographic change alone should be interpreted in terms of either 
Continental Drift or long-distance dispersal; but Occam’s Razor should discourage 
the invocation of a long-distance dispersal hypothesis until or unless theories based on 
normal dispersal distances prove to be untenable. 

(iii) Long-distance-dispersal theories, when they have been proposed, should be related to 
methods of dispersal. 

Using the above criteria, the only section in which evidence of long-distance dispersal has 
been found in Hypericum is sect. 30. Spachium, where it appears to have occurred in several 
groups among the most specialised members, all of which grow in damp habitats. 


3. Morphology 
Habit 


Species of Hypericum exhibit a wide range of habit forms from the arboreal to the 
ephemeral. At one extreme, some species in sect. |. Campylosporus (e.g. H. bequaertii) can 
form trees up tol2 m high, with a true single trunk; at other extremes one finds, for example, 
H. gentianoides (sect. 30. Spachium), an annual herb with leaves reduced to scales, and H. 
elodes (sect. 28. Elodes), which can produce threadlike stems in water up to 50cm deep 
(Gliick, 1911). 

True trees, however, are rare and occur only among the most primitive species of the most 
primitive sections (1. Campylosporus, 3. Ascyreia, 29. Brathys). Most of the woody species 
are many-stemmed, with the main branches erect or suberect. The laterals may spread to 
some extent; but only in relatively advanced species does one find all the branches spreading 
with co-planar leaves. This effect can be so pronounced as to give a frondose appearance to 
the whole shoot (e.g. in H. uralum, sect. 3. Ascyreia). The lowermost branches in these 
shrubs may arise from below soil level, but they do not normally root, even in the dwarf 
shrubs such as H. olympicum (sect. 10. Olympia) which are on the border of woodiness. 

Among perennial herbs, however, the situation is different. Many of these spread at or 
near ground level by means of runners, which grow horizontally for some distance before 
becoming erect or ascending, and the horizontal nodes frequently produce roots. In H. elodes 
this habit has become very well developed, as it has in some other species of wet habitats (e.g. 
H. scioanum, sect. 30. Spachium). In other species a true rhizome is found (e.g. in H. 
adpressum, sect. 20. Myriandra). At least two species also produce vegetative buds from the 
roots, H. perforatum (sect.9. Hypericum) and H. pulchrum (sect. 18. Taeniocarpium), a 
phenomenon which may partly explain the abundance of these species in western Europe 
(Salisbury, 1942). The successful establishment of H. perforatum after introduction into 
other parts of the world, however, is more likely to relate to its methods of seed production. 

The annual species, which of course all have perennial ancestors, do not show this 
propensity for rooting at the nodes. The transition may be observed in H. humifusum 
(sect. 14. Oligostema), for example, in which the perennial forms tend to have rooting shoots 
whereas the annual ones (‘H. /iottardii Vill.’) have only the tap-root system. 


Stems 
(a) Stem-lines and phyllotaxis 


The co-ordination of leaf-initiation in Hypericum is always complete, so that the phyllotaxis 
is always either opposite-decussate or rarely 3-4-whorled. In this respect it resembles all the 
other genera of the Guttiferae except most members of the Bonnetioideae and some species 
of Psorospermum Spach. In this genus the phyllotactic co-ordination breaks down in the 
upper parts of the stem of some species (e.g. P. alternifolium J. D. Hook.), resulting in 
secondarily alternate leaves. 

Associated with the opposite-decussate phyllotaxis is the occurrence of four raised lines on 
the younger parts of the stem in many species (Fig. 8a). This appears to be the primitive 


74 N. K. B. ROBSON 


Fig. 8 Stem lines in Hypericum and Ploiarium: (a) Hypericum revolutum; (b) H. perforatum; (c) 
H. bithynicum; (d) H. nitidum; (e) Ploiarium alternifolium. 


state in Hypericum; and it gives rise to 2-lined and then terete conditions, either by 
development (on the same woody shoot) or evolution (in some herbaceous species) (Fig. 8b, 
c). Occasionally six lines may be present in some internodes, e.g. in some species of 
sect. 20. Myriandra (Fig. 8d). Although these structures are usually no more than ‘raised 
lines’, they can be wide enough to be termed ‘wings’ (e.g. in H. tetrapterum, sect. 9. 
Hypericum). The members of an opposite pair are always equally prominent; but there is 
almost always some difference between alternating pairs so that the lines decurrent from the 
leaf-bases (medians) are more prominent than those from between the leaves (laterals). 
Consequently, where there is only one pair of lines, these will almost always be medians. The 
exceptions to this rule occur in sect. 5. Androsaemum, where the sometimes 4-lined H. 
hircinum has laterals that are more prominent than the medians and the 2-lined H. 
androsaemum has laterals only. 

The association of stem-lines with leaves is clearer in Ploiarium alternifolium (Bonneti- 
oideae), in which the presence of stem-lines in the youngest parts is combined with spiral 
phyllotaxis. Although the lines are sometimes faint, three can be seen below each leaf, 
decurrent respectively from the midrib and each side of the petiole (Fig. 8e). It would 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 75 


appear, then, (i) that the stem-lines are related morphologically to the leaves, (ii) that the 
opposite-decussate phyllotaxis in Hypericum usually results in the lateral lines uniting in 
pairs (4-lined stems), but that (iii) they occasionally remain distinct (6-lined stems). In the 
occasional species or forms with 3-4-whorled leaves (e.g. H. empetrifolium, sect. 19. 
Coridium; H. ternatum, sect. 23. Triadenioides; H. ternum, sect. 30. Spachium), the number 
of lines equals the number of leaves, i.e. only medians are present. 

Evidence produced by Noack (1939) from crosses between H. perforatum L. (2-lined, 
tetraploid) and H. maculatum subsp. maculatum (4-lined, diploid) (sect. 9. Hypericum) 
indicates that the presence of lateral lines in this group is an incompletely dominant 
character, the triploid hybrids (H. x desetangsii Lamotte nm. carinthiacum (Frohl.) N. 
Robson) having very narrow or incomplete laterals. 


(b) Woody stems 


In the shrubby and arborescent species the internodes eventually become terete, although the 
lines may remain in evidence for a considerable time (e.g. in Hypericum revolutum, 
sect. 1. Campylosporus). The young stems may be green or vinous red, but the secondary 
ones are smooth (sometimes even ‘polished’) and reddish or brownish-red to cinnamon at 
first, without any visible lenticels. They may eventually become almost black. In most 
species vertical fissures then appear and the cork layers turn pale grey and flake off in strips. 
Sometimes, however, for example in H. frondosum and H. lissophloeus (sect. 20. 
Myriandra), these vertical fissures do not appear and the cork layers exfoliate in larger 
patches, as in species of Betula. 

In A. chapmanii (also sect. Myriandra), Adams (1962a) describes a soft spongy bark up to 
4 cm thick, of which the outer layers disintegrate and fall away, leaving more-resistant strips 
of hardened laticifers. Other species in this section (e.g. H. tetrapetalum) have less spongy 
bark, and in other sections it rarely becomes markedly corky. 

The stems can eventually attain a diameter of 15 cm (in H. chapmanii) or more (in some 
species of sects 1. Campylosporus and 29. Brathys). 


(c) Stem glands 


The epidermis of herbaceous stems in Hypericum sometimes contains small, round or some- 
what elongated glands similar to those occurring in the leaves (cf. pp. 80-82). These may 
be colourless, reddish or black, depending on the concentration of hypericin present in them, 
and they may be confined to the stem lines (sect. 9. Hypericum) or dispersed over the whole 
stem (sects 12. Origanifolia, 17. Hirtella). In sect. Hirtella the glands vary in colour and 
prominence. Thus, in H. /ydium they range from colourless to reddish, and the reddish ones 
may be distinctly prominent (‘H. adenocladum’ Boiss.). This last trend is more marked in the 
typical variety of H. scabrum, in which the reddish stem glands are on short unbranched 
protuberances, giving the stem the roughness that suggested the epithet “scabrum’; and in H. 
thymopsis, which is derived from H. scabrum, the gland-tipped protuberances toward the 
base of the stem are branched. 


\Vestiture 


Apart from glandular emergences, the only vestiture that occurs in Hypericum consists of 
simple uniseriate hairs. No stellate hairs, such as those that are a feature of the Vismieae, are 
found. Members of the primitive sections are wholly glabrous or have a minutely papillose 
leaf epidermis (e.g. H. revolutum subsp. keniense (Schweinf.) N. Robson, sect. 1. Campylo- 
sporus) (cf. Spirlet, 1967), but an indumentum has developed independently in several 
advanced groups of species. Thus, species in sects 1-10 and 13-16 are wholly glabrous, 
except for the Bulgarian H. setiferum (sect. 13. Drosocarpium), which differs from H. 
montbretii only in having sparse appressed hairs on the lower surface of the leaf and is 
probably not specifically distinct from it. The stem, leaves and sepals of various species in 


716 N. K. B. ROBSON 


sects 11. Campylopus, 12. Origanifolia, 17. Hirtella and 18. Taeniocarpium are scabrid- 
papillose to hirsute, but the hairs remain distinct and spreading. Likewise, in the only species 
in sect. 30. Spachium with indumentum (H. setosum), the hairs are distinct and relatively 
short. 

All the remaining hirsute species are found in sect. 27. Adenosepalum and its ‘satellite’ 
sect. 28. Elodes, and among them only in species of northern Africa, Macaronesia, Europe 
and western Asia. In H. reflexum, from Macaronesia, only the stem is pubescent, and the 
European and N. African H. montanum is wholly glabrous or has a scabrid lower surface of 
the leaf. Otherwise, where hairs are present they are distributed over the stem, leaves and 
sepals (outer surface only), and they sometimes become long and interwoven (e.g. in H. 
tomentosum and H. elodes). 


Leaves 
(a) Insertion and margin 


Like most genera of the Bonnetioideae, the primitive species of Hypericum have sessile 
leaves. Where a petiole has evolved, it is never completely differentiated from the lamina, 
the transition being always gradual. Leaves with a relatively long petiole of this type occur in 
sects 9. Hypericum (e.g. in H. petiolulatum) and 27. Adenosepalum (e.g. in a form of H. 
himalaicum N. Robson), but species in several sections can be described as sub-petiolate. At 
the other extreme, six distinct species or species-groups with perfoliate leaves have evolved, 
viz. H. bupleuroides (sect. 8. Bupleuroides), H. sampsonii (sect. 9. Hypericum), H. spectabile 
(sect. 17. Hirtella), H. pamphylicum (sect. 22. Arthrophyllum—cf. Robson & Davis, 1980), 
H. caprifolium and H. coadunatum (sect. 27. Adenosepalum) and H. connatum and H. 
caprifoliatum (sect. 30. Spachium). 

Associated with changes from shrubby to herbaceous habit there is a tendency towards 
retention of the leaves on the stem. Primitive species, therefore, have an articulation at the 
base of the leaf, whereas in advanced species this has disappeared (cf. Adams, 1962a). 

The most primitive species in sects 1. Campylosporus and 29. Brathys have the leaf-base 
reflexed, suggesting an auricle, and the same structure occurs in Santomasia steyermarkii. In 
more advanced species of Hypericum, however, this structure has been lost; but true auricles 
have evolved in three sections, in each of which they are associated with gland-fringed leaf 
margins, thus providing the only exceptions to the rule that the leaves in the Guttiferae 
(excluding some Bonnetioideae) are entire. In two sections (viz. 15. Thasia and 16. Crosso- 
phyllum, i.e. in H. thasium, H. orientale and H. adenotrichum), the evolution of gland- 
fringed leaves and auricles seems to have been basipetal from the bracts, as auricles are less 
conspicuous in or absent from the lower leaves. In H. vesiculosum (sect. 13. Drosocarpium), 
on the other hand, it is the lower leaves that are gland-fringed, whereas the upper ones are 


entire and there is an abrupt transition to the narrow gland-fringed bracts (see also Bracts 
and bracteoles, p. 85). 


(b) Venation 


All species of Hypericum have a complete midrib, and the most primitive ones (H. 
bequaertii, H. revolutum subsp. keniense and H. lanceolatum subsp. angustifolium (Lam.) 
N. Robson—sect. 1. Campylosporus and H. styphelioides—sect. 29. Brathys) have ‘parallel’ 
venation. In order to understand the evolution of leaf venation within the genus, however, it 
is necessary to consider in more detail some members of allied genera that were discussed 
briefly in relation to Santomasia (p. 61). 

In the Bonnetioid Neblinaria celiae Maguire (1972 : 155, f. 24) (Fig. 9a), the crowded 
leaves have a flat pseudopetiole which is entered by about seven leaf traces. These 
dichotomise considerably in the broad base of the leaf before radiating without further major 
branching to near the apex. The main veins are equally thick apart from the median one, 
which is somewhat thicker than the rest at the base and forms an embryonic midrib. In other 
leaves (e.g. that shown in Fig. 9a), there is a tendency for the veins adjacent to this midrib to 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 


| 


uae tH Ss 
Si fie AE 7 “a S 
eae dia | | SS 
Cn] 1] H 

(i hi \ LARP | 

ta ie | 


Fig. 9 Leaf-venation in some Bonnetioideae and Hypericeae, showing evolutionary trends: (a) 
Neblinaria celiae (x 1); (b) Neogleasonia wurdackii (x 1); (c) Neotatea colombiana (x 1); (d) 
Santomasia steyermarkii (x 2); (e) Hypericum bequaertii (x 2); (f) Hypericum styphelioides (x 5). 


77 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 79 


unite with it. Near the apex the main veins may dichotomise and anastomose to form a 
reticulum, but the margin itself is unveined. As we have seen above (p. 61), this venation is 
basically pedati-flabellate or parallelodromous, but the marginal anastomoses towards the 
apex show a tendency towards the development of an intramarginal vein. 

Neogleasonia wurdackii Maguire (1972 : 160, f. 25), (Fig. 9b), also Bonnetioid, has less 
crowded leaves which do not narrow at the base. The leaf scars show one large central trace 
(which becomes a complete midrib) and four small lateral ones. Some or all these appear to 
dichotomise in the leaf base, as in Neblinaria; but the tendency for the veins nearest the 
midrib to become united with it is more marked, so that, in effect, the midrib is pinnately 
branched in the lower third or more. Another development from Neblinaria is that the veins 
have differentiated in ‘size’ so that thick ones alternate with thin ones. Neogleasonia, 
however, does not have marginal reticulation like Neblinaria. 

In the third of Maguire’s new Bonnetioid genera, Neotatea (Fig. 9c), the lateral veins are 
all decurrent into the midrib and the submarginal anastomoses have been consolidated, 
resulting in a densely pinnate venation (paxillate—Melville, 1976) with an intramarginal 
vein. 

In Santomasia (Fig. 9d) the first of these trends is more advanced than in Neogleasonia, in 
that the process of union with the midrib has continued further. There are 6-10 lateral veins, 
of which the lower 4-5 are decurrent into the midrib and the rest (which are less distinct from 
the tertiary venation) join it at a definite, wider angle. As regards the number of its lateral 
veins, S. steyermarkii is more primitive than Hypericum bequaertii (Fig. 9e); but in-so-far as 
these are decurrent into the midrib, and in its densely reticulate tertiary venation and its 
glandularity (see p. 61), it is far more advanced. If one takes the numerous veins and 
incomplete midrib of Neblinaria as the primitive state, then S. steyermarkii is more 
advanced in having (a) fewer veins, (b) a complete midrib and (c) the veins united with the 
midrib. 

The leaf-trace in Hypericum and Santomasia is unilacunar, as it is in all the Guttiferae so 
far investigated (including the Bonnetioideae) (Schofield, 1968). Maguire (1972 : 134) stated 
that the nodal anatomy of the Bonnetioid genera might not be consistent; but it seems 
reasonable to assume that they, too, are unilacunar, even though Neogleasonia and Neotatea 
have 3-trace leaf-scars (in Neblinaria they are 1-trace). If one accepts the hypothesis that the 
paired unilacunar trace is primitive in the angiosperms (cf. Schofield, 1968), then the 
incomplete midrib in Neblinaria can be conceived as derived by the partial union of the 
innermost branch of each trace. 

Hypericum bequaertii is advanced relative to Santomasia in the trend towards a reduction 
in the number of veins. Thus, in H. bequaertii (Fig. 9e) there are only 3-5 major veins. On 
the other hand, these veins sometimes arise from dichotomies in the leaf-base (i.e. they are 
not always united with the midrib); and the leaf is at the Neogleasonia stage in having 5-7 
thinner ‘veins’ alternating with the major ones. 

Although these weaker ‘veins’ are apparently part of the same branching structure as the 
midrib and major basal veins, their connections with it are often absent. Thus one can some- 
times follow such veins back through several dichotomies to the base of the leaf but find no 


Fig. 10 Trends in major leaf-venation in Hypericum (numbers indicate sections): (a) H. 
bequaertii (1); (b) H. revolutum subsp. keniense (also H. lanceolatum subsp. angustifolium) (1); 
(c) H. revolutum subsp. revolutum (also H. lanceolatum subsp. lanceolatum) (1); (d) H. 
madagascariense (1); (e) H. balearicum (2); (f) H. canariense (21); (g) H. cardiophyllum (22); (h) 
Hi. quartinianum (1); (i) H. roeperanum (1); (j) H. synstylum (1); (k) H. frondosum (20); (1) H. 
glandulosum (27); (m) H. lanuginosum (27); (n) H. elodes (28); (0) H. socotranum (1); (p) H. 
scopulorum (23); (q) H. lysimachioides (17); (r) H. hirsutum (18); (s) H. olympicum (10); (t) H. 
orientale (16); (u) H. bithynicum (13); (v) H. choisianum (3); (w) H. androsaemum (5); (x) H. 
elatoides (7); (y) H. pseudopetiolatum (9); (z) H. brasiliense (30); (a) H. silenoides (30); (6) H. 
mutilum (30); (¢) H. styphelioides (29); (a) H. magniflorum (29); (é) H. phellos (29); 
(f) H. stenopetalum (29); (g) H. struthiolifolium (29); (A) H. brathys (29). N.B. For amended 
position of f, g (sects 21, 22), see footnote on p. 187 and Fig. 3, p. 68. 


80 N. K. B. ROBSON 


clear connection with the main venous system. Indeed, they are not always continuous and 
have apparently lost (or are losing) their conductive function to become secretory structures. 
The conductive system has therefore been reduced to the midrib and two or four main basal 
or near-basal veins, none of which are branched. In other primitive species, e.g. H. 
quartinianum (sect. 1. Campylosporus) or H. styphelioides (sect. 29. Brathys) (Fig. 9f), there 
may be up to eight main lateral veins, but the venation is advanced in other respects relative 
to that of H. bequaertii in that these lateral veins are branched. 

With the proviso, then, that H. bequaertii may be specialised relative to some other species 
in the number of its basal veins, it may be taken as the basic species from which to show 
trends in major leaf-venation in the genus (Fig. 10a). These trends involve (i) adnation of the 
basal veins to the midrib (e.g. in H. quartinianum, H. hirsutum) (Fig. 10h, r), (11) pinnation of 
the midrib (e.g. in H. revolutum, H. brasiliense) (Fig. 10b, c, z), (i11) incurving of vein ends to 
form an intramarginal vein (e.g. in H. roeperanum, H. frondosum) (Fig. 10i, k), (iv) conden- 
sation of the lower part of the midrib, so that the main lateral veins radiate from one point 
(e.g. in H. elodes, H. mutilum) (Fig. 10n, b’), (v) elimination of some or all basal veins so that 
eventually only an intramarginal vein remains (e.g. in H. balearicum, H. scopulorum) (Fig. 
10e, p) and (vi) reticulation, i.e. the in-filling of spaces between main veins or laterals with 
increasingly dense reticulate venation (e.g. in H. androsaemum, H. frondosum) (not shown 
in Fig. 10). Two further points relating to Fig. 10 should be noted: (i) reduction trends result- 
ing in the elimination of all veins except the midrib occur in other sections besides 29. 
Brathys (e.g. in sects 17, 18, 26, 30), but only where an intramarginal vein has not formed; (11) 
an apparent increase in the number of basal veins is sometimes (? always) due to branching 
(e.g. in some plants of H. elatoides) (Fig. 10x). 


(c) Glandularity 


The glandular pattern in the leaves is, in part, intimately connected with that of the 
venation, and so it is appropriate to discuss it here. 

There are two types of gland present in the leaves of Hypericum: (i) Blackish or red clusters 
of cells containing a wax impregnated with hypericin (a dianthrone) or occasionally pseudo- 
hypericin, and (ii) schizogenous intercellular spaces lined by cells which secrete a translucent 
essential oil (Coutinho, 1950; Mathis, 1963; Mathis & Ourisson, 1963, 1964a, b). These can 
conveniently be called ‘dark’ and ‘pale’ glands, respectively. Pale glands are present in all 
species (Green, 1884), although they may not be visible externally in some heavily sclero- 
tinized or cutinized leaves; but dark glands are sometimes completely absent from the plant 
or confined to organs other than the leaves, and this variation in occurrence has been used in 
the classification of the genus (Clos, 1868; Siersch, 1927; Roth, 1953; Mathis & Ourisson, 
1963) (see p. 149). 

The pale glands are also of two types that have quite distinct evolutionary origins: 

(i) As we have shown already, the main veins in H. bequaertii, H. revolutum subsp. 
keniense and H. lanceolatum subsp. angustifolium are interspersed with weaker, dichoto- 
mising, apparently functionless veins (Figs 9e, 1la) which appear to have evolved by 
differentiation from an equal-veined flabellate condition (Fig. 9). With branching of the 
midrib and secondary veins the course of these functionless veins (which have become linear 
‘pale glands’) has been interrupted (Fig. 11b), and formation of increasingly denser tertiary 
reticulate venation has resulted first in glandular streaks and then isodiametric glands e.g. in 
H. roeperanum or H. revolutum subsp. revolutum (Figs 9, 11). 

(ii) Interspersed with the veins and vein-glands in H. bequaertii are irregular rows of small 
pale glandular dots (Fig. 11a). These glands do not seem to correspond to any structures 1n 
other subfamilies of the Guttiferae, but the glands in Santomasia and the Vismieae appear to 
be homologous with them. The pale glands in the Cratoxyleae, however, seem to have had 
an origin like those in type (i) of Hypericum. Glands of type (ii) are present in sect. 1. 
Campylosporus only in H. bequaertii, H. revolutum subsp. keniense and H. lanceolatum 
subsp. angustifolium. It is the only type to occur in sects 29. Brathys and 30. Spachium, the 


—————— 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 81 


ran ravage, 
Veena aoe: 


i 
WE} 
i, 


Fig. 11 One origin of pale glands in leaves of Hypericum: (a) H. bequaertii (x 6); (b) H. revolutum 
subsp. keniense (x 12); (c) H. revolutum subsp. revolutum (x 12); (d) H. roeperanum (x 4). 


HOHE 


BRC. vat 


rans pi Poot nn q SS Bets 


[el 


functionless veins having apparently disappeared completely during the evolution of these 
sections or, in Brathys, perhaps they never evolved; and glands of this type are present in 
some species of sect. 3. Ascyreia as small, superficial dots. All other pale glands in the leaves 
of Hypericum are derivatives of the functionless veins (i.e. type (1)). 

The dark glands are associated particularly with the leaf margins and the region 
immediately interior to them, and Green (1884) pointed out that they were ‘in intimate 
relationship with the fibrovascular system’. This is true of only the more primitive species. 
In other parts of the genus there are ‘laminar’ as well as ‘marginal’ dark glands; and these 
develop in positions which, in related species, are occupied by pale glands. There is no 
doubt, however, that Green’s observation is partially valid; and a comparison with the 
Bonnetioideae confirms its validity. In Ploiarium alternifolium (Vahl) Melchior and 
Bonnetia stricta (Nees) Nees & Martius, for example, the leaf margins have indentations 
towards the apex each containing a sessile gland or gland-tooth which terminates a vascular 
strand. In the leaves of species of Hypericum sect. 1. Campylosporus there are comparable 
glands, but these do not interrupt the contour of the margin; they are thus marginal or (when 
more deeply imbedded) inframarginal or submarginal. In H. bequaertii they are amber in 
colour and therefore contain no hypericin, only the wax base; but in some other species (e.g. 
H. roeperanum) they are reddish or dark due to their hypericin content. 

As was mentioned above, this variability in the strength and occurrence of hypericin can 
be used to characterise sections or large parts of the genus (Fig. 12; Robson, 1977a: fig. 1). 
For example, in sect. 3. Ascyreia and some of its derivatives (sects 4, 5, 7), red or black glands 
are totally absent. In sects 6. Jnodorum and 8. Bupleuroides they are sometimes present on 
sepal margins only, and then usually only reddish. In the Hirtella group (sects 17-19) they 
are usually only marginal in the perianth and usually either absent from the leaves or apical 
or laminar, not marginal. They may, however, occur on the stem. In the Olympia group 
(sects 10-16) they are marginal and often also laminar in the leaves and perianth but do not 
occur on the stem; whilst in sect. 9. Hypericum they are similarly distributed but do also 
occur on the stem. 


82 N. K. B. ROBSON 


Ss) Black glands absent 


25 | 24 | ce 14 | i (ae Only reddish glands present 


Fig.12 Variation in distribution of black or reddish glands in Hypericum. 


On the other hand, related species may differ widely in the occurrence of hypericin. In 
sect. 27. Adenosepalum black glands are numerous and widespread in the more primitive 
species; but in the apparently derived sect. 28. Elodes the glands are usually merely reddish 
and confined to the sepal margins. A similar diminution in hypericin content is shown in the 
eastern Himalayan members of sect. 9. Hypericum, which are derived from a group of 
Chinese species with dark glands. An even more vivid example is found in sect. 16. 
Crossophyllum, where, of the two species (both with gland-fringed leaves), H. adenotrichum 
has dark marginal and superficial glands, whereas in H. orientale all the glands are pale (not 
even reddish). In the Vismieae, on/y amber to dark glands are present, whereas the 
Cratoxyleae have only pale glands. 


Inflorescence 
(a) Form 


As in the Bonnetiaceae, the inflorescence in Hypericum is basically cymose; but it never has 
a truly axillary position, i.e. in the axils of foliage leaves with the terminal bud remaining 


Fig. 13 Trends in inflorescence-branching in Hypericum (numbers indicate sections): (a) H. 
bequaertii (1); (b) H. revolutum (1); (c) H. aegypticum (25); (d) H. heterophyllum (24); (e) H. 
quartinianum (1); (f) H. roeperanum (1); (g) H. cordifolium (3); (h) H. scabrum (17); (i) 
salsolifolium (17); G) H. punctatum (9); (k) H. brasiliense (30); (1) H. laxiusculum (30); (m) H. 
myrianthum (30); (n) H. styphelioides (29); (0) H. chamaemyrtus (30); (p) H. diosmoides (30); (q) 
H. carinosum (29); (r) H. thuyoides (29); (s) H. natalense (26); (t) H. peplidifolium (26); (u) H. 
tortuosum (23). O Flower, @ bud. 


84 N. K. B. ROBSON 


vegetative (as in Ploiarium or Neogleasonia). The most primitive inflorescence in 
Hypericum is 1|-flowered, as it is in Kielmeyera, for example; and the opposite decussate 
phyllotaxis results in elaboration by formation of dichasia or pseudo-dichotomies. These 
dichasia, sooner or later and to differing degrees in various sections, give way to monochasia. 
Reduction can, of course, set in at any evolutionary stage, the dichasia being replaced by 
single flowers. Although the branching is basically cymose at the level of the inflorescence, 
so that laterals do not develop until the terminal flower is initiated, in the young plant of e.g. 
H. revolutum (sect. 1) they appear long before that event. The main shoot at that stage is 
therefore indeterminate (hapaxanthic); and the ‘architecture’ of the branching would seem to 
conform with the type described as Scarrone’s Model by Hallé, Oldeman & Tomlinson 
(1978): 

‘The architecture is determined by an orthotropic rhythmically active terminal meristem 
which produces an indeterminate trunk bearing tiers of branches, each branch-complex 
orthotropic and sympodially branched as a result of terminal flowering.’ 


The terminal bud (always?) eventually forms a flower, and then the ‘single trunk’ form 
gives way to a ‘bush’ form. Development of shoots from the base of the main stem (i.e. 
basitony sensu Hallé, Oldeman & Tomlinson, 1978) is common and contributes to the 
aspect of many species. 

The basic state of branching at the flowering level is well shown in H. bequaertii (Fig. 
13a). Below the terminal solitary flower are one or more nodes in which branch develop- 
ment is absent, presumably having been inhibited by hormones from the apical meristem. 
Along the remaining length of stem a varying number of branches are produced, the most 
advanced ones being somewhere in the middle (i.e. immediately below the inhibitory zone). 

Two of the three major trends in elaboration of the inflorescence are discernible in 
sect. 1. Campylosporus: (1) In e.g. H. revolutum and H. madagascariense the central, 
branching zone becomes continuous (1.e. all nodes have branches), but the branches remain 
1-flowered (Fig. 13b). (2) In e.g. H. quartinianum, on the other hand, the uppermost node(s) 
produce single flowers or triads (Stauffer, 1963), there being usually a non-branching zone 
between these and the uppermost flowering branches (Fig. 13e). In H. roeperanum this non- 
branching zone is absent and there is a continuous progression from terminal flower, through 
lateral flowers, triads (1-noded inflorescence branches) and paracladia (2-or-more-noded 
inflorescence branches), to the uppermost flowering branch (where the inflorescence module 
is repeated on a reduced scale, in this example reduced to a single flower) (Fig. 13f). The 
paracladia may bear triads only, or the subterminal node(s) may also be fertile. (3) The third 
major trend in inflorescence elaboration can be described as pseudo-dichotomy (Fig. 13n), 
in that both axillary buds of the terminal node (below the flower) produce branches, not 
dichasia. If the flower falls, as it does in H. styphelioides, for example, the resultant effect is of 
repeated dichotomized branching. The occurrence of pseudo-dichotomy is confined to only 
four sections: 20. Myriandra, 26. Humifusoideum, 29. Brathys and 30. Spachium. 

All Hypericum inflorescences can thus be described in terms of modification of (i) the 
terminal flower and the nodes immediately beneath it and (ii) the intermediate zone of 
1-flowered branches, i.e. whether the terminal elaboration is dichasial or pseudo- 
dichotomous, and whether or not the intermediate zone is present. To these two modifi- 
cations, which may be termed respectively acrotonal and intercalary, must be added a third 
(basitonal), in that the nodes occupied by the inflorescence proper (i.e. terminal flower, 
dichasia and paracladia) may increase in number basipetally (Figs 13c, e) until eventually 
the non-branching zone is eliminated (Fig. 13f). Where this has occurred, it may be 
impossible to differentiate paracladia from lateral branches (cf. Stauffer, 1963). Basitony 
combined with reduction of the dichasia to single flowers results in a spike, e.g. in Z. 
geminiflorum (sect. 4. Takasagoya) or H. salsolifolium (sect. 18. Hirtella) (Fig. 131). 

Dichasial to monochasial branching is the only type of acrotony that occurs in sect. 3. 
Ascyreia and its derivatives (sects 4-19), and also in sects 20-25, 27. Adenosepalum and 
28. Elodes (Fig. 13c-j). In some species (e.g. in sect. 30. Spachium, Fig. 13m), dichasia of up 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 85 


to the 4th order may be produced before monochasium-formation sets in. In others (e.g. 1n 
sect. 9. Hypericum, Fig. 13j) it sets in after the appearance of the 2nd-order flowers. 

Dichasium-formation is rare in sect. 29. Brathys (occurring, for example, in H. phellos, H. 
pseudocaracasanum and H. humboldtianum), and the dichasia are never elaborate. The 
primitive type of branching in this section is pseudo-dichotomy (e.g. in H. styphelioides, Fig. 
13n); and, where this is the sole type of branching (e.g. in H. bryoides), a dense rounded 
shrub is formed, approximating to Leeuwenberg’s Model (Hallé, Oldeman & Tomlinson, 
1978). In this section, further evolution has mainly been by elaboration of the intercalary 
zone (e.g. in H. carinosum or H. phellos, Fig. 13q), leading eventually in H. /aricifolium and 
its relatives to retarded development of the inflorescence proper (i.e. the terminal flower) and 
the conversion of the shoot from determinate to indeterminate growth (Fig. 13r). In sects 
20. Myriandra, 26. Humifusoideum and 30. Spachium both types of branching are found, 
sometimes in the same inflorescence (‘mixed inflorescence’) (Fig. 130, p, s). 


Another cause of differences in inflorescence size, though not in complexity, is variation in 
the length of the internodes. Extreme condensation produces a subumbellate corymb, e.g. in 
H. scabrum (sect. 17. Hirtella) (Fig. 13h); elongation may result in a candelabra-like form 
(e.g. in H. monogynum and its relatives, sect. 3. Ascyreia). 

Extreme reduction can result in either the elimination of branches, producing a second- 
arily 1-flowered inflorescence (e.g. H. nummularium, sect. 19. Taeniocarpium, or forms of 
H. anagalloides, sect. 30. Spachium), or sterilisation (i.e. delayed development) of lateral 

branches (as in H. heterophyllum, sect. 24. Heterophylla) (Fig. 13d). In some species (e.g. H. 
‘peplidifolium, sect. 26. Humifusoideum; H. elodes, sect. 28. Elodes; H. scioanum, sect. 30. 
Spachium), development of only one of the pair of axillary buds at the terminal node results 
in sympodial branching and apparently lateral inflorescences (Fig. 13t). As regards 
elaboration, inflorescence branches always arise singly, except in H. /axiusculum (sect. 30. 
Spachium), where accessory branches occasionally occur at the terminal node (cf. St Hilaire, 
1824-1828 : t. 61). 

In conclusion, which method of acrotonal branching seems to be ‘the most primitive’? In 
most parts of the genus the elaboration is in the order: single flower—dichasium— 
-monochasium, or, where mixed inflorescences occur: single flower—pseudo-dichotomy— 
dichasium—monochasium. In sect. 29. Brathys, however, the single-flower stage is rare and, 
as stated above, dichasium-formation is rarely present and then rudimentary. Nevertheless, 
we should conclude that, here too, the pseudo-dichotomous inflorescence is derived. 


(b) Bracts and bracteoles 


In Hypericum bequaertii the transition in form from leaf to sepal is sudden, the uppermost 
leaves being only a little smaller than the others; and wherever the inflorescence is 
1-flowered, the same tends to be true. For example, most species of sect. 29. Brathys show 
this sudden transition, only a few having upper leaves intermediate in form. 

In more complex inflorescences and their derivative forms, however, the bracts (upper- 
most leaf pair) and bracteoles tend to be more similar to the sepals than to the leaves, i.e. the 
sudden transition is often from the leaves to the bracts. This resemblance to the sepals 
extends to glandularity. In general, the bracts and bracteoles are persistent for at least as long 
as the leaves, i.e. where the leaf-base is articulated (see p. 76), the bracts and bracteoles tend 
to be deciduous; but in sects 3. Ascyreia and 4. Takasogoya they are caducous. 

In sects 15. Thasia and 16. Crossophyllum, and also, independently, in sect. 27. 
Adenosepalum, the basipetal extension of sepaline characters has penetrated beyond the 
inflorescence into the foliage leaves. In some African species of sect. Adenosepalum (e.g. H. 
annulatum), the sepals and bracts are gland-fringed, and at the base of the bracts the 
glandular cilia form auricle-like clusters; whereas, in the most closely related Asian species 
(H. elodeoides, H. himalaicum, H. wightianum), the upper leaves also usually have 
glandular fringes and glandular auricles. Likewise, in sect. Thasia the upper leaves have 


86 N. K. B. ROBSON 


glandular fringes and auricles; but in sect. Crossophyllum all the leaves are gland-fringed and 
have true auricles, not merely clusters of glandular cilia (see p. 76). 


Sepals 
(a) Insertion and number 


The transition from the opposite decussate phyllotaxis of the leaves to the 2/5 spiral 
(quincuncial) phyllotaxis of the calyx appears to be relatively sudden in most species of 
Hypericum, in that the upper leaves and bracts are usually opposite. Occasionally, however, 
the bracteoles or upper leaves are not strictly opposite, e.g. in H. tomentosum (Wydler, 
1851), or one of the bracts may be absent (Wydler, 1859, 1871). This tendency towards 
merging of the phyllotactic systems never becomes a permanent feature in Hypericum, 
whereas in some west African species of Psorospermumi (P. alternifolium J. D. Hook., P. 
guineense (L.) Hochr., P. corymbiferum Hochr. and P. lanatum Hochr.) the upper leaves and 
branches are constantly alternate to spiral. A comparison of these species with the rest of the 
Vismieae indicates that this is an advanced state, not one transitional from the spiral phyllo- 
taxis typical of most Bonnetioideae and the Theaceae to the opposite decussate arrangement 
typical of the rest of the Guttiferae. 

The suddenness of the transition from opposite to quincuncial insertion in Hypericum is, 
however, often more apparent than real. In the large, regularly pentamerous flowers of 
sect. 3. Ascyreia, the two outermost sepals (1 and 2) do not diverge by the ideal 2/5 angle of 
144°, but by one that averages 180°. Breindl (1934) records a range of 160°-200° for ‘H. 
hookerianum’. Likewise, sepals 3 and 4 are only about 20° divergent from the opposite 
position in this species (161°-199° in one flower measured). It is only by adding together the 
divergence angles of each pair that space can be provided for sepal 5, which diverges from 
sepal 3 by only 54’, not by the ideal 72°. These spatial relationships are reflected in temporal 
ones, in that the almost opposite pairs of sepals arise nearly simultaneously, i.e. 1-2, 3-4, 5 
(cf. discussion below under 6. Development, p. 125). In flowers with meiomerous inner 
whorls the angle between sepals | and 2 tends to become larger, and sepals 3, 4 and 5 then 
arise together, i.e. the order is 1-2, 3-4-5. 

These trends in spatial and temporal relationships of the sepals explain how there is a 
tendency towards crowding in the inner floral whorls, with consequent elimination of the 
gynoecial members on the radii of sepals 5 and frequently 4, and the fusion of androecial 
members on either side of these radii (see Fig. 21, p. 104). This tendency occasionally 
extends to the perianth in sections where this is normally pentamerous, e.g. in H. humifusum 
(sect. 14. Oligostema) (Martin-Sans, 1922) or in some populations of H. monanthemum 
Hook.f. & Thomson ex Dyer from Sikkim and Nepal, which were described as Ascyrum 
filicaule by Dyer (1874).* The tendency towards tetramery is most marked, however, in 
sect. 20. Myriandra, where several species have a constantly tetramerous perianth and 
androecium and comprise the Linnaean ‘genus’ Ascyrum (Robson, 1977a, 1980c). Tran- 
sitional conditions are found in this section, for example in H. frondosum, where the sepals 
are often foliaceous and the calyx and corolla may vary from pentamerous to tetramerous on 
one plant. In all the typically tetramerous species except H. microsepalum (which is not 
closely related to them, see Figs 4, 5), the elimination of sepal 5 has been followed by 
differentiation of the remaining ones into a larger outer pair and a smaller inner pair, these 
pairs being of course at right angles to one another. In H. hypericoides the smaller, inner pair 
is sometimes obsolete (i.e. vestigial) (fide Adams 1957). 

An example of an apparently opposite trend is found in the prostrate, few-flowered form of 
H. empetrifolium (sect. 19. Coridium) that is known horticulturally as var. prostratum 
(correctly var. oliganthum K. H. Rech.). In it the influence of the 3-whorled phyllotaxis has 


*In the first paper of this series (Robson, 1977a), Hypericum /filicaule (Dyer) N. Robson was considered to be distinct 
from H. monanthemum. Further work has shown these taxa to be conspecific and to belong to sect. 9. Hypericum, 
not to sect. 27. Adenosepalum. 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 87 


been carried over to the perianth, which is hexamerous. In both calyx and corolla the 
irregular disposition of the members indicates that an extra member has been ‘added on’. 
There is no evidence of true trimery. Wydler (1878) found comparable variation in the 
flowers of H. coris, another species of sect. Coridium. 


(b) Relative size, shape, union and persistence 


The difference in size between inner and outer sepals in species with a tetramerous perianth 
(e.g. Hypericum hypericoides) is an extreme manifestation of a general tendency in parts of 
the genus towards inequality in sepal size—a tendency that is associated with the spatial and 
temporal relationships discussed above. The most primitive condition, as shown in H. 
bequaertii and its near relatives, is slight inequality, the inner sepals being usually somewhat 
smaller than the outer ones, with sepal 3 intermediate in size. 

In some sections (e.g. 17. Hirtella) there is a tendency towards equality (e.g. in H. 
scabrum), whereas in others the tendency is in the opposite direction. Foliaceous or very 
unequal sepals have evolved repeatedly, e.g. in sect. 3. Ascyreia (H. calycinum, H. prattii, H. 
choisianum), sect. 5. Androsaemum (H. androsaemum), sect. 7. Roscyna (H. przewalskii) 
and sect. 14. Oligostema (H. humifusum). The evolution of markedly unequal sepals is 
usually accompanied by a tendency towards their spreading or reflexing in bud and/or fruit, 
but these developments are associated with no obvious biological changes related, for 
example, to pollination or dispersal. 

Sepal shape is extremely variable and shows no general trends, although narrowness is 
usually relatively advanced. There is, however, a trend towards union of the sepals, which 
are primitively quite free at the base. In the most extreme forms, e.g. in H. cardiophyllum 
(sect. 22. Arthrophyllum), H. kotschyanum (sect. 18. Taeniocarpium) or H. scabrum 
(sect. 17. Hirtella), the sepals may be up to 2/3 united; but in most species they are free or 
united only at the base. Union is not particularly associated with the development of pseudo- 
tubular flowers and specialised insect pollination (in sects 25. Adenotrias and 28. Elodes), as 
the sepals are free in the former and 2/5 united in the latter. Whether united or not, the 
imbricate quincuncial arrangement is always visible in bud, although in calyces with 
relatively narrow sepals the aestivation at the base may be almost open, e.g. in H. brachy- 
phyllum (sect. 20. Myriandra). 

In most Hypericum species the sepals persist during and after development of the fruit; but 
in two separate sections they are sometimes deciduous. In sect. 5. Androsaemum they are 
persistent in all except H. hircinum, even in its hybrid with H. androsaemum (H. x 
inodorum Miller). In sect. 20. Myriandra, however, the sepals are primitively deciduous 
after fruit dehiscence (i.e. in subsect. Myriandra) but become persistent again in all but one 
of the derivative herbaceous or tetramerous species (subsect. Pseudobrathydium R. Keller) 
(cf. Adams, 1962a). 


(c) Vestiture 


Only the outer surface of the sepals ever has an indumentum, and that resembles the vesti- 
ture of the corresponding vegetative parts. There are no species in which hairs are confined 
to the sepals; but in H. reflexum and H. annulatum (sect. 27. Adenosepalum) or H. setosum 
(sect. 30. Spachium), for example, the vegetative parts are wholly or partly indumentum- 
covered, whereas the sepals are glabrous. 


(d) Venation 


| The sepals in Hypericum are basically 3-trace organs (see p. 129). The three traces may 


depart from the same stelar gap (Fig. 32, p. 129); but more usually there are three gaps, with 
_the laterals of adjacent sepals (often united initially) departing from the same gap (Fig. 35, p. 
131). The three traces then dichotomise several times in the base of the sepal or not, 


depending on its width, so that the sepal has one to many veins, parallel or diverging (Fig. 


88 N. K. B. ROBSON 


14). In the more primitive and many relatively advanced species, further branching is also 
dichotomous (e.g. H. roeperanum, sect. 1. Campylosporus, Fig. 14a), but pinnation often 
occurs, especially towards the margin (e.g. in species of sect. 9. Hypericum, where it is 
accompanied by the looping of the laterals on to the midrib or on to each other, Fig. 14b). In 
foliaceous sepals (e.g. in H. przewalskii, Fig. 14c) a densely reticulate venation is produced; 
but this is always clearly a development from parallel or diverging veins, unlike the reticu- 
lation in the leaves, which tends to develop from midrib branches (cf. Fig. 10). Thus, in 
general (and for any given species) the venation of the sepals is at a less advanced stage of 
evolution than is that of the leaves. Where the sepals are narrow, however, the laterals and 
branchings may be few or absent, as in narrow leaves (Figs 10h, 14d). 


Fig. 14 Variation in sepal venation in Hypericum: (a) H. roeperanum; (b) H. sampsonit; (c) H. 
przewalskii, (d) H. brathys (all x 6). 


(e) Margin 


Like the sepal margins of the Bonnetioid genera, that of Hypericum is primitively entire (Figs 
15a, 16); but sect. 1. Campylosporus contains species in which it is gland-fringed (Fig. 1 5b), 
and even in H. bequaertii it sometimes has minute eglandular denticles. These do not appear 
to be associated with vein-endings. In some sections the margin remains constantly entire 
(e.g. in 5. Androsaemum, 20. Myriandra, 29. Brathys) (Fig. 15u, y), whilst in others it is con- 
stantly fringed (e.g. in 13. Drosocarpium, 19. Coridium) (Fig. 15h, i); but in several sections 
both conditions are found (e.g. in 9. Hypericum, 17. Hirtella) (Fig. 15e, r), and the change 
from entire to fringed seems to have occurred independently in most of them (Fig. 16). 
Although this change is usually associated with the emergence of marginal glands, in some 
sections an eglandular eroded-denticulate margin has developed (e.g. in 3. Ascyreia, 9. 
Hypericum, 30. Spachium) (Fig. 15c, f). In general, an entire sepal margin is a primitive 
character, but there has been a reversion from gland-fringed to entire sepals in at least four 
sections: 9. Hypericum (H. monanthemum), 11. Campylopus (H. cerastoides), 14. 
Oligostema (H. australe, H. humifusum pro parte) and 27. Adenosepalum (e.g. H. 
himalaicum N. Robson=H. humifusum sensu Dyer non L.) (cf. Fig. 15j, k). In some 
sections (e.g. 1. Campylosporus) the denticles precede the glands (i.e. in these sections they 
are primitively eglandular), whereas in others (e.g. Webbia) (Fig. 15v) the emergences are 
glandiferous from the beginning. 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 89 


Ea 


a 3 


Fig. 15 Variation in marginal contour and in glandularity of sepals in Hypericum (numbers 
indicate sections): (a) H. bequaertii (1); (b) H. quartinianum (1), (c) H. forrestii (3); (d) H. 
formosanum (4); (e) H. erectum (9); (f) H. maculatum subsp. obtusiusculum (9); (g) H. 
polyphyllum subsp. polyphyllum (10); (h) H. montbretit (13); (i) H. barbatum (13); Q) H. 
linarifolium (14); (k) H. humifusum (14); (1) H. thasium (15); (m) H. orientale (16); (n) H. 
hirtellum (17); (0) H. salsolifolium (17); (p) H. uniglandulosum (17); (q) H. retusum (17); (r) H. 
asperulum (17); (s) H. pumilio (18); (t) H. fragile (18); (u) H. prolificum (20); (v) H. canariense 
(21); (w) H. elodeoides (27); (x) H. wightianum (27), (y) H. strictum (29); (z) H. brasiliense (30) 
(all x 5). 


90 N. K. B. ROBSON 
(f) Glandularity 


The sepals bear the same types of gland as do the leaves, viz. (1) black or red cell-clusters 
containing hypericin and (ii) schizogenous pale punctiform or elongate lacunae containing 
essential oils. The occurrence and form of these glands vary in different parts of the sepal, 
which can be divided into four regions: (i) laminar—not associated with the margin; (ii) 
submarginal—towards the margin; (iii) inframarginal—immediately below the margin but 
not breaking the line of it; (iv) marginal—breaking the line of the margin. The various types 
of gland are homologous with those in similar positions in the leaves, the bracts and 
bracteoles, as already stated, being intermediate in glandularity between leaf and sepal. 


Fig. 16 Distribution of entire sepal margin in Hypericum (within line). 


(i) Laminar. As was observed above with regard to venation, so it is with laminar glands: 
the sepals in general show a less advanced state than do the leaves, i.e. they retain pale 
glandular lines (canals) for longer. Thus, in a given species the chances are that, if the 
laminar pale glands of the leaf are dots or streaks, those of the sepals are at least partly linear. 

Although these glandular lines occur between the veins, as in the primitive leaves, never- 
theless they do not appear to have a relation to functionless veins. Nor is their dissection 
apparently related to increasingly reticulate venation, as the lines are frequently discon- 
tinuous towards the margin when the veins are not or scarcely branched. There is neverthe- 
less a trend from almost unbroken lines* in the primitive sections (and Santomasia) to 
increasing distal and marginal dissection, culminating in sepals with punctate or shortly 
elongate glands only, e.g. in Hypericum thasium (sect. 15. Thasia) (Fig. 151) and H. 
wightianum (sect. 27. Adenosepalum) (Fig. 15x). 

The dark laminar glands, like those of the leaf lamina, occur in positions which corre- 
spond to those of pale glands in related species or sections. Thus, whereas the sepals in many 


*Where the sepals are thick, the lines, though continuous, may appear broken (cf. Fig. 14a). 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 91 


species of sect. 17. Hirtella have two pale linear glands, in H. retusum these are replaced by 
two dark linear glands, and in H. asperulum they are dark streaks or dots (Fig. 15q, r). The 
replacement may be incomplete, so that a linear gland is mostly pale but has dark sections 
(e.g. in forms of H. erectum, sect. 9. Hypericum or H. elodeoides, sect. 27. Adenosepalum) 
(Fig. 15e, w). Although direct replacement of pale line by dark line can therefore occur, the 
trend in dark glands is more usually from dots to lines by means of linear extension or 
coalescence. 


(ii) Submarginal. In many species the dark glands towards the margin are punctiform 
whilst those nearer the midrib and base are longer. But a few species have a definite row of 
dark dots or short streaks which, while not touching the margin, are associated with it. These 
seem to be confined to sect. 1. Campylosporus (Hypericum revolutum, H. gniidfolium, H. 
lanceolatum, H. madagascariense, H. quartinianum, H. roeperanum) (Fig. 15b), although in 
H. polyphyllum (sect. 10. Olympia) they approach the submarginal position (Fig. 15g). As 
they are merely dark distal portions of otherwise pale glands, these submarginal glands are 
really only a special type of laminar gland. 


(iii) Inframarginal. The inframarginal glands, on the other hand, are a special type of 
marginal gland that is associated with entire sepals, whereas marginal glands (by definition) 
occur on an interrupted sepal margin. Indeed both types may occur in the same species (e.g. 
Hypericum humifusum) (Fig. 15k). In at least two species (H. thasium, sect. 15. Thasia and 
H. wightianum, sect. 26. Adenosepalum) inframarginal and stipitate marginal dark glands 
appear to occur on the same sepal, both species having sepals with glandular or eglandular 
fimbriae between which are dark glandular dots that break the line of the margin (Fig. 151, 
x). By comparison with related species (respectively H. montbretii and H. elodeoides) (Fig. 
15h, w), however, it is clear that these are laminar glands that appear to have become 
marginal due to the deep incision of the margin. 

Inframarginal sepal glands, like inframarginal leaf glands, are always punctiform or 
elliptic, never more elongated. 


(iv) Marginal. Whereas the laminar and submarginal glands obviously develop between 
the veins, there has been disagreement about the relative position of the marginal and 
inframarginal ones. Clos (1868) stated that the ‘marginal’ (i.e. non-stipitate) glands occur 
between the veins, whilst the glandular cilia appear at the ends of the veins; but Green (1884) 
described both as being ‘in intimate relationship with the fibrovascular system.’ My observ- 
ations lead me to agree with Clos that the association with veins occurs only in stipitate 
glands. But it does not always occur. In species where the marginal glands are primitively 
small (and often reddish rather than dark, e.g. in Hypericum revolutum, sect. 1. Campylo- 
sporus or H. canariense, sect. 21. Webbia (Fig. 15v)), there is no association with the 
vascular system. Nor is there any vascular connection with the eglandular marginal denticles 
in, for example, H. forrestii (sect. 3. Ascyreia) and H. maculatum subsp. obtusiusculum 
(Tourlet) Hayek (sect. 9. Hypericum) (Fig. 15c, f). It would seem, therefore, that the 
marginal accumulation of glandular material is the primary phenomenon, the association 
with the vascular system being a secondary one. 

The trend from inframarginal via marginal to stipitate glands is reversible. Thus, in 
sect. 14. Oligostema, H. linarifolium has subequal sepals with ciliate margins comparable 
with those of the most closely related species in sect. 12. Origanifolia—and indeed with most 
species in sects 12-16 (Fig. 15j). The cilia mostly contain a vascular strand. In the very 
closely related H. australe and H. humifusum (Fig. 15k), the sepals tend to be entire and 
(particularly in H. humifusum) unequal, when the glands revert to the inframarginal position 
with no vascular connection. 

The distribution of dark glands, as we have already seen, can be used as a taxonomic 
character (p. 80; Robson, 1977a). In some sections the sepals have no dark glands (e.g. 2-5, 
7, 20, 24, 25, 29, 30), in some they are small and rudimentary (6, 8, 21), in others they are 
sometimes present (1, 22, 23), and in the remainder they are constantly present (9-19, 


92 N. K. B. ROBSON 


26-28). The dark glands may occur alone, but dark laminar glands are nearly always 
accompanied by dark marginal ones. An exception to this rule occurs in sect. 10. Olympia, 
where the entire sepals of H. polyphyllum are without marginal glands but usually have 
laminar dark glands (Robson, 1967a, b, 19805) (Fig. 15g). 

Marginal dark glands first appear either as small, sessile, reddish or blackish, globose 
structures (sects 6, 8, 17, 18, 21) or as minute amber to blackish denticles (sect. 1). In more 
advanced stages of evolution they may terminate glandular cilia (short, spreading, not 
interrupting the marginal line), glandular denticuli (short, ascending, not interrupting the 
marginal line), glandular fimbriae (long, spreading, not interrupting the marginal line) or 
glandular laciniae (long, spreading or ascending, interrupting the marginal line). In some 
more advanced species the emergences become eglandular (e.g. in H. barbatum, sect. 13. 
Drosocarpium and H. thasium, sect. 15. Thasia) (Fig. 151, 1), or the glands may lose some or 
all of their hypericin and become red (e.g. in H. elodes, sect. 28. Elodes) or amber (e.g. in H. 
orientale, sect. 16. Crossophyllum) (Fig. 15m). The shape of the stipitate glands also varies. 
They are usually globose or ellipsoid; but obconic ones occur in sects. 9. Hypericum (e.g. H. 
monanthemum), 17. Hirtella (e.g. H. hirtellum) (Fig. 15n) and 27. Adenosepalum (e.g. H. 
annulatum); and, at least in sect. Hirtella, this difference in shape can be of taxonomic 
importance (Robson, 1967c, 1968). 

What appears to be a special type of marginal dark gland occurs in sects 17. Hirtella and 
18. Taeniocarpium, sections in which such glands are normally absent from the leaves. In 
sect. Hirtella, the upper leaves and sometimes the sepals of H. salsolifolium (Fig. 150) have a 
subapical or apical more-or-less irregular accumulation of dark glandular material, the 
apical mucro of the leaf being eglandular; whilst in the closely related H. uniglandulosum 
(Fig. 15p), where both leaf and sepal are mucronate, the subapical gland is absent, but the 
mucro is often glandiferous. In sect. Taenocarpium, the leaves (but not the sepals) of H. 
nummularium have two conspicuous apical dark glands, which in the taxonomically 
adjacent H. monadenum have united to form one large gland. Other species of this affinity 
have 1-5 small apical leaf glands (H. pumilio) (Fig. 15s) or none (H. nummularioides, H. 
fragile, H. saxifragum) (Fig. 15t). In the taxonomically somewhat isolated H. haplophyl- 
loides, an Albanian endemic species which comprises two disjunct populations, whereas 
both subspecies have apical dark glands in the leaf, only one (subsp. devollense F. K. Meyer) 
has sepals with a glandiferous apex (Meyer, 1978). 


Petals 
(a) Insertion and number 


Whereas the phyllotaxis of the leaves and the perianth of the Dilleniaceae and Theaceae are 
constantly spiral, the Guttiferae show transitions to whorled phyllotaxis in both regions. But, 
whilst the transition in leaf-insertion is apparent in the Guttiferae only in the Bonnetioideae, 
as we have seen (p. 73), that in the perianth is more widespread. In transitional stages of 
perianth aestivation, instead of having two organs wholly exterior, two wholly interior and 
one (no. 3) half exterior and half interior (Fig. 17a), there are fewer wholly exterior or 
interior and more in the half-and-half position characteristic of the contorted or convolute 
state (Fig. 17b-d). Thus, in the Bonnetioideae the aestivation is intermediate (species of 
Bonnetia and Neotatea) or contorted, in the Vismieae it is always quincuncial, and in the 
Cratoxyleae it is quincuncial (Eliea, Cratoxylum), intermediate (Eliea, Cratoxylum, 
Triadenum) or contorted (Cratoxylum, Thornea, Triadenum) (own observations). 

In Santomasia and Hypericum, irrespective of whether the petals are 4, 5 or (abnormally) 
6 in number, the aestivation is always contorted. Wydler (1878) describes the direction of 
twist in H. elegans as continuing that of the calyx spiral, and this appears to be the rule in the 
genus. The number, like that of the sepals, is reduced to 4 in those species of sect. 20. 
Myriandra that have been treated as the genus Ascyrum L., as well as in H. filicaule 
(sect. 9. Hypericum) and abnormally in some usually 5-merous species. Transitional states 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 


o~ van oo~ a 
a = ome: 


Fig. 17 eta aestivation in pentamerous members of the Guttiferae: (a) quincuncial (as in 
Theaceae); (b, c) intermediate forms (as in Triadenum fraseri); (d) contorted (as in Hypericum). 


VS 


can be observed in H. frondosum (sect. 20. Myriandra) (see p. 198). The numbers of 
members in these two whorls are almost always the same; only very exceptionally has an 
extra petal been recorded (e.g. in H. papuanum, sect. 26. Humifusoideum, and H. empetri- 
folium, sect. 19. Coridium). 


(b) Form 


A consequence of the constantly contorted aestivation of the petals in Hypericum is that, ina 
given flower, the same margin of each petal is outside in bud. It may be this relationship that 
has resulted in the development of asymmetry in the petals, a character that is absent in all 
allied genera except Santomasia (and Cratoxylum, where it is sometimes slightly developed). 
This asymmetry is always evident to some extent, except (i) in the smaller, reduced flowers 
(e.g. of H. humifusum, sect. 14. Oligostema or H. japonicum, sect. 30. Spachium) and (ii) in 
some flowers with a pseudo-tubular corolla (sects 25. Adenotrias, 30. Elodes). In other, less 
highly evolved species the inner and outer margins are always distinguishable. 

The line of the outer margin is frequently interrupted by a notch or indentation, which 
results in a more or less evident projection. This marks the limit of the part of the margin that 
is actually outside in bud and has been termed the apiculus (Robson, 1970). Even when an 
apiculus is not visible, the corresponding point on the margin (which is the morphological 
apex) is usually recognisable from the pattern of venation and glands. It marks the end of the 
midrib, which divides the petal into two, usually unequal, parts: (i) an outer, usually thicker 
and smaller part with denser venation and glands more often laminar, and (ii) an inner, 
usually larger and thinner part with laxer venation and glands more often marginal. 

Hypericum petals are nearly always eligulate, as are those of the Bonnetioideae, most of 
the other Hypericoideae and all the rest of the Guttiferae. Development of a ligule has 
occurred in two separate sections of Hypericum (25. Adenotrias, 28. Elodes) and in Psoro- 
spermum, Cratoxylum (sects Tridesmos and Isopterygium) and Eliea (Perrier de la Bathie, 
1951; Gogelein, 1967; Baas, 1970; Robson, 1974); and in all these taxa except Cratoxylum 
sect. Isopterygium and Psorospermum (where the appendages are very small), this develop- 
ment is associated with the evolution of a pseudo-tubular flower and specialised insect 
pollination (Robson, 1972a, 1974, 1977a). In all but Hypericum sect. Elodes, Psoro- 
spermum and Cratoxylum sect. Isopterygium, the presence of a ligule is also associated with 
dimorphic heterostyly; but this association is not essential for the development of heterostyly 
in the Hypericoideae, as some species of Vismia (with petals internally villous) are hetero- 
styled and have no petal ligule (Robson, 1978). 

In Hypericum sect. 25. Adenotrias the ligule is entire, concave and cucullate; and it is 
pressed closely against the ovary, thus leaving a gap opposite each sepal. In sect. 28. Elodes, 
however, it is flat and trifid, and the association with the ovary is less close. In both sections 
the ligule is long, non-vascular and united with the petal lamina for most of its length. 

These appendages have been termed nectary scales, but they do not appear to secrete 
nectar. Their function may be purely structural (associated with the pseudo-tubular corolla), 
or there may be juicy cells between them and the petal lamina which are pierced by insect 


94 N. K. B. ROBSON 


tongues. This question has not yet been investigated anatomically. What seems clear, how- 
ever, is that they have the same morphological status as petal ligules in, for example, the 
Ranunculaceae or Amaryllidaceae. It remains to be discovered whether the ligules in the 
other Hypericoid genera are non-vascular like those of Hypericum. At any rate, the primitive 
petal in the subfamily (as in the Bonnetioideae) would appear to be eligulate. 


(c) Colour 


In the Guttiferae sensu lato the petal colour varies from white and yellow or green to shades 
of red; blue is entirely absent. Almost the whole of this range of colour is represented in the 
Hypericoideae, but the tribes of this subfamily are more limited in this regard. The Vismieae 
have shades of white, yellow and green; in the Cratoxyleae yellow is replaced by red, the 
petals being white or pink to red, sometimes tinged with green; whilst in the Hypericeae 
green is absent and the basic petal colour is normally a shade of yellow, often with red tinges 
or markings. Rarely in Hypericum is the basic colour white, and then the tinges and 
markings are pink. This effect appears to be due to the red anthocyanin’s becoming paler 
when the basic flavonoid is white instead of yellow. 

The yellow basic (flavonoid) colour in Hypericum varies from orange (aurantiacus), e.g. in 
H. bequaertii and several species in sect. 30. Spachium, through dark yellow (luteus) and 
golden yellow (aureus) to pale yellow (flavus). Occasionally an even paler shade (citrinus) 
occurs as a ‘sport’, e.g. in H. olympicum forma minus Hausskn. c.v. Sulphureum (sect. 10. 
Olympia) (Robson, 19805) and H. hirsutum var. pallidum Rouy & Foucaud (sect. 18. 
Taeniocarpium) (Druce, 1923; Brenan, 1945). Pure white forms are very rare but have been 
recorded in H. geminiflorum (sect. 4. Takasagoya), from Luzon (Merrill, 1909 : 294, as H. 
loheri) and probably Taiwan (Robson, 1976), and H. aviculariifolium (sect. 12. Origanifolia) 
(Robson, 1967a, b). This colour variation is rarely of taxonomic significance above the 
species level, but certain generalisations can be made. Thus, the deeper shades (apricot, 
orange) occur only in sects 1. Campylosporus, 29. Brathys and 30. Spachium; deep yellow is 
more widespread (e.g. in Santomasia and H. acmosepalum, sect. 3. Ascyreia) but is not 
confined to a particular section; in sect. 9. Hypericum the petals tend to be golden yellow, 
whereas in sect. 27. Adenosepalum they are usually paler in colour. 

The red colour (anthocyanin) usually occurs in the outer dorsal part of the petal, either 
confined to the veins (e.g. in H. trichocaulon, sect. 13. Drosocarpium) or as a more or less 
extensive tinge. In only one taxon has it extended over the whole corolla (and calyx), 
resulting in a ‘red-flowered Hypericum’ (H. capitatum var. capitatum, sect. 17. Hirtella); but 
other members of the same species (H. capitatum var. luteum N. Robson) have no red 
colour in these organs (Robson, 1967a, b). Again, whereas H. tetrapterum, like most of 
sect. 9. Hypericum, has no petal anthocyanin, its sister-species H. undulatum frequently has 
red-tinged petals. They are wholly yellow only in var. boeticum (Boiss.) Lange, which is 
intermediate in form and distribution between the type variety and H. tetrapterum. 

The presence of anthocyanin in other parts of the Hypericum plant is often an indication 
of an unfavourable environment; and this association may be relevant with regard to H. 
capitatum, which grows in a habitat that approaches desert more nearly than does that of any 
other Hypericum. But the function of the red tinges has been shown to be related to insect 
visitation (see p. 122). 


(d) Venation 


In common with the rest of the Guttiferae and the other families in the Dilleniidae, 
Hypericum and Santomasia have |-trace petals (Saunders, 1937; Robson, 1956, 1972a) (Fig. 
32, p. 129). In larger petals (e.g. those of H. forrestii, sect. 3. Ascyreia) the subsequent basal 
dichotomies form a slender midrib and two thicker traces which, in turn, very soon dicho- 
tomise repeatedly to form a series of veins of roughly equal width (Robson, 1972a: fig. 3) 
(Fig. 18b). In smaller petals the branchings are fewer, and the small petals of some herbs in 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 95 


d e 


Fig. 18 Variation in venation and glandularity of petals in Hypericum (numbers indicate 
sections); (a) H. styphelioides (29); (b) H. forrestii (3); (c) H. perforatum (9); (d) H. asperulum 
(17); (e) H. elodeoides (27) (a, b x 2°5;c x 3°5;d x 4; ex 5). 


sect. 30. Spachium have only three, unbranched veins. The veins are crowded and little- 
branched in the morphologically outer part of the petal but more widely spaced and more 
frequently dichotomising in the inner part (Fig. 18a).* Towards the margin they sometimes 
form loops or lambda junctions (Melville, 1969, 1976) with adjacent veins (Fig. 18b—d); but 
they frequently reach the margin unbranched (Fig. 18b) or (sometimes excepting the midrib) 
stop short of it without forming a loop or junction (Fig. 18e). The evolutionary progressions 
in general seem to be: 

Veins numerous few; much-branched unbranched; free—*looped or joined; all 
reaching margin—+only midrib reaching margin—+none reaching margin. 


(e) Glandularity 


The petal glands in Hypericum are comparable with those of the sepals, except that truly 
inframarginal glands are rare. When the glands form a distinct peripheral row, they are dark 
and usually touch or interrupt the margin in some way. Such a row is present in the petals of 
Santomasia in the form of glandular cilia or sessile glands, sometimes along with a few larger 
submarginal dark glands (Fig. le). In all other genera of the Hypericoideae, as well as those 
of the Bonnetioideae, the petal margin is always entire and marginal glands are absent. Their 
presence in the Hypericeae may therefore be regarded as an advance. Once acquired, 
however, they can be lost again (e.g. in sect. 3. Ascyreia) and then reacquired (e.g. in 


*In Archytaea (Bonnetioideae), on the ofher hand, the petal is symmetrical and the veins are equal and symmetrical 
also, cf. Maguire (1972 : fig. 2d). 


96 N. K. B. ROBSON 


sect. 9. Hypericum); and so their presence is taxonomically significant at the sectional level 
at most. 


(i) Laminar glands. In Santomasia and primitively in Hypericum the crowded veins 
towards the base of the petal alternate with pale glandular lines, which are always inter- 
rupted distally to some extent; but in the thick petals of these taxa, the lines may be easily 
visible only towards the base, and then incompletely as a series of dots (Figs le, 18a, b). The 
inner, thinner part always remains eglandular (Fig. 18c). The interruptions become 
gradually more numerous to form streaks and eventually dots (e.g. in sects 17, 18 and 19) 
(Fig. 18d). They are sometimes much reduced in number as well as size; but they never dis- 
appear entirely in a section where they are present in the primitive members. They do, how- 
ever, seem to be absent entirely from sects 25. Adenotrias and 30. Spachium. 

In addition to this variation in form and distribution, the petal laminar glands may become 
dark or reddish. The whole gland may be dark (e.g. in H. elodeoides, sect. 27. Adenosepalum 
(Fig. 18c), or only parts of a glandular line may be dark and the rest pale (e.g. in H. 
oliganthum, sect. 9. Hypericum). 


(ii) Marginal glands. The presence of truly marginal petal glands in the Bonnetioideae— 
Hypericoideae affinity is confined to the Hypericeae, and it is probably a derived character 
in this tribe. In Hypericum it has become constant in the less primitive members of sect. 1. 
Campylosporus, in two groups derived from sect. 3. Ascyreia (the Olympia group, 1.e. sects 
10-16, and sect. 17. Hirtella), and in some members of sects 9. Hypericum, 18. Taenio- 
carpium and 27. Adenosepalum. It therefore appears that the occurrence of marginal petal 
glands in Hypericum is polyphyletic, and it possibly indicates a certain minimum concen- 
tration of hypericin in the petals (see p. 149). Their presence in Santomasia and sect. 1. 
Campylosporus, respectively, is to be regarded as a local specialisation, a trend that is 
reversed in sects 27. Adenosepalum and 28. Elodes. Here the gradual diminution in the 
concentration of hypericin is reflected in the petal marginal glands, which are present in the 
primitive (Macaronesian) species but absent from the most advanced ones. 

The marginal glands may be sessile (e.g. in H. revolutum subsp. keniense, sect. 1. 
Campylosporus) or on cilia (e.g. sometimes in H. revolutum subsp. revolutum). Where they 
occur in the Hirtella group (sects 17-19) they are usually on cilia, whereas they are nearly 
always sessile (and often in a marginal depression) in the Olympia group (sects 10-16) and 
sect. 9. Hypericum, except in H. sampsonii (sect. 9. Hypericum), where they are on cilia. 
Where they do occur they are nearly always confined to, or are more strongly developed on, 
the inner margin; and they are always dark except in H. orientale (sect. 16. Crossophyllum) 
and some advanced members of sect. 27. Adenosepalum, where they are reddish or amber. In 
both these groups, as has already been mentioned, there would seem to have been a 
diminution in the concentration of hypericin. 


(f) Corolla shape 


In general, Hypericum corollas are radiate, i.e. the petals spread out or become reflexed. The 
only notable exceptions are: (i) where the sepals remain erect and confine the petals, forming 
a pseudo-tubular subrotate corolla (sects 25. Adenotrias, 28. Elodes) and (11) in H. bequaertii, 
where the petals remain erect so that the corolla is campanulate. It is not clear whether this 
form is primitive in Hypericum or derived. On the one hand, erect petals occur in 
Santomasia and in primitive genera of the Moronobeoideae (Pentadesma Sabine and 
Montrouziera Planchon & Triana), as well as in the Vismieae and Cratoxyleae. On the other 
hand, radiate corollas are almost universal in the rest of the Guttiferae (including the 
Bonnetioideae). As H. bequaertii seems to have differentiated from the H. revolutum subsp. 
keniense (with a radiate corolla) in response to peculiar climatic conditions of the high 
Ruwenzori (Robson, 1979), perhaps the campanulate corolla is best regarded as a local 
specialisation. 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 97 
(g) Corolla movement 


With the above-mentioned exceptions, the petals in Hypericum spread more or less widely 
or become reflexed. This movement may occur once (e.g. in sect. 3. Ascyreia) or may be 
repeated daily for a few days (e.g. in sect. 10. Olympia). In some species of sect. Ascyreia, 
however, the concave petals do not become reflexed but form a shallowly cupuliform corolla 
(compare the reflexed petals of H. kouytchense with the ascending ones of H. beanii in 
Robson, 1970: figs 235, 240). The extremely reflective surface of the petals in this section 
would produce the effect of a paraboloid mirror in a cupuliform corolla and might concen- 
trate the sun’s rays to some extent, raising the temperature inside the flower. 


(h) Corolla (and androecium) persistence 


One of the features used by Keller (1893, 1925) in classifying Hypericum was whether or not 
the petals and stamens were deciduous immediately after anthesis, and this has proved to be 
a useful character at sectional level (Robson, 1977a: fig. 1). In general the petals and stamens 
behave similarly in this regard, which is not surprising when one considers that petal and 
stamen fascicle together form an androphyll (Melville, 1963) and are supplied by the same 
vascular trunk (Robson, 1956, 1972a). The only exception to this rule is found in all but two 
species of sect. 20. Myriandra subsect. Pseudobrathydium, in which the stamens remain 
when the petals fall (Adams, 1962qa) (Fig. 5). It may be significant that this change in 
behaviour has occurred in the one section where, the individuality of the fascicles having 
been completely lost, the stamens in the primitive species are shed individually, not as 
fascicles. 

A similar development seems to have occurred in the Bonnetioideae. In Ploiarium, where 
the androecium is obviously fasciculate, both petals and stamens are deciduous; whereas in 
the other genera, where the androecium is outwardly afasciculate, the stamens are persistent 
and the petals deciduous. It is not yet clear whether deciduous or persistent petals are 
primitive in the Guttiferae as a whole. From first principles, deciduous ones would be 
derived; but, as we have seen, the change seems to be easily reversible and would not there- 
fore appear to be fundamental. In Hypericum this is certainly so (Fig. 19). In sect. 1. 
Campylosporus both petals and stamens are persistent; but they have become deciduous in 
three adjacent sections or groups, viz. sect. 2. Psorophytum, sects 3-5 (H. socotranum is tran- 
sitional) and sect. 20. Myriandra (H. quartinianum and H. synstylum are transitional). In 
two further sections, 19. Coridium and 25. Adenotrias, the change has been intrasectional. 
Reversion to persistent stamens has taken place in sect. 20. Myriandra; and persistent petals 
and stamens are constant in all the sections derived from 3. Ascyreia (i.e. sects 4-19) with the 
exception of 4. Takasagoya, 5. Androsaemum and the above-mentioned part of 19. 
Coridium. 

The biological significance of deciduous floral members in Hypericum is not clear. Where 
all outer members fall, as in H. hircinum (sect. 5. Androsaemum) or the primitive members 
of sect. 20. Myriandra, the fruit is naked. Where the sepals remain, as in sect. 3. Ascyreia, 
they have no obvious function (e.g. of protection). Where all whorls remain for at least some 
time after the corolla and androecium have withered, the inner ones may behave in two 
distinct ways. In H. olympicum and most species of the Olympia and Hirtella groups (sects 
10-19) they twist up round the developing fruit, forming a bud-like protective covering. In 
other species (e.g. sect. 7. Roscyna and most of sect.9. Hypericum) the withered petals 
remain spreading or reflexed and twist or curl, and the stamens become suberect. Although 
the behaviour of the corolla is characteristic of whole groups of species, it must be used 
taxonomically with care. For example, although in most of the Olympia group the withered 
whorls become bud-like, in some species of sect. 13. Drosocarpium (e.g. H. perfoliatum) they 
remain spreading. In contrast, although in many species of sect. 9. Hypericum they remain 
spreading, in H. undulatum they become bud-like. There may be a correlation between the 
relative width of the petals and their behaviour; I can think of no other explanation of this 
phenomenon. 


98 N. K. B. ROBSON 


Petals and stamens persistent (inside line) 
” " deciduous (outside line) 


---=Petals deciduous, stamens persistent 


Fig.19 Distribution of deciduous petals and stamens in Hypericum. 


Stamens 
(a) Number and arrangement 


The association between petals and stamen fascicles mentioned above suggests that the 
fascicle of stamens, rather than the individual filament + anther, is the androecial unit in 
Hypericum, and the correlation of five free antipetalous fascicles with other primitive 
characters in sects 1-4 makes this hypothesis more likely. In contrast, to suggest that fascicles 
have resulted from ‘dédoublement’ or ‘secondary polyandry’ (Moquin-Tandon, 1826; Payer, 
1857; Saunders, 1936; Leins, 1964, 1971) or from aggregation of stamens or sectorial steri- 
lisation in a polyandrous flower (Goebel, 1898-1901; Hirmer, 1917; Corner, 1946) makes 
nonsense of other trends in the genus (Robson, 1956, 1972a). The only species which some- 
times have five single antipetalous stamens (H. gentianoides and H. aphyllum Lundell, | 
sect. 30. Spachium) are annual herbs with leaves reduced to scales; and the apparently 
polyandrous species (in sects 20. Myriandra, 29. Brathys and 30. Spachium) are all special- | 
ised in several respects in comparison with those of sect. 1. Campylosporus. The stamen | 
fascicle in Hypericum thus appears to be a reduced branching system such as is envisaged in } 
the Telome Theory (Zimmermann, 1959; Wilson, 1937, 1953) and Gonophyll Theory . 
(Melville, 1963). 

If the stamen fascicles form the antipetalous androecial whorl then the antisepalous whorl 
is normally absent in Hypericum. In Santomasia, however, as we have already seen, the 
latter is present as small sterile projections between the stamen fascicles (Fig. 1g). Compari- 
son with related taxa shows that the diplostemonous pentamerous androecium with each 
member represented by a multistaminate (usually fasciculate) structure should be regarded 
as primitive in the Dilleniidae (cf. Wilson, 1965, 1973; Stebbins & Hoogland, 1976). In some 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 99 


orders (e.g. Malvales, Violales) it is the antipetalous whorl that is reduced or absent (van 
Heel, 1966); in others it is the antisepalous one (e.g. Theales, Myrtales*), whilst in still others 
(e.g. Ericales) both are equally reduced—to one-stamened members. A transitional state to 
the Malvalean one is found in Saurauia (Saurauiaceae), where Brown (1935) showed that the 
apparently polyandrous androecium develops as five single antipetalous stamens and five 
groups of antisepalous ones. In the Guttiferae sensu lato both whorls are sometimes present, 
but the only genus in which they are both known to be fertile is Decaphalangium Melchior 
(1930) from Amazonian Peru. It seems likely, however, that anatomical investigation will 
show that both whorls are present in at least some species of Clusia and its relatives. In other 
tribes the antipetalous whorl is sterile or absent (Table 3). 


(b) Modifications of fascicles 


If the fascicle is a reduced branching system, then this reduction has resulted in the virtual 
elimination of the primitive fascicle trunk above toral level. There would appear to be no 
extant species of the Dilleniales—-Theales evolutionary line in which the stamen fascicle has a 


Table3 Characters of the androecium in bisexual and male flowers of the Guttiferae 


Taxon Antisepalous whorl Antipetalous whorl* 
Presence and No. and union No. and union Union of 
fertilityt of fascicles of fascicles filaments 

Bonnetioideae sora 5 5 or (5) free 
Hypericoideae 
Hypericeae sora 5 (Santomasia), 0, 5, (5), (4), free or 
3 (Hypericum (2)+14+1+41, partly 
sects 25, 28) (2) +(2)+1 united 
(Hypericum fe 
sects 25, 28, 30 pp.) 
Vismieae S 5 5 partly united 
Cratoxyleae S 3 (2)+(2)+1 partly united 
Moronobeoideae S 5, (5) 5,16) free to 
wholly 
united 
Calophylloideae a 0 (5), (4) free to 
wholly 
united 
Clusioideae 
Clusieae f(? ora) 5, (5) (? or 0) 5 (5) free to 
wholly 
united 
Decaphalangium f 5 5 wholly 
united 
Allanblackia S 5 5 wholly 
united 
Garcinieae sora 5, (5), 4, (4) 5, (5), 4, (4), 3 free to 
wholly 
united 


*The antipetalous whorl is always fertile, except in 9 flowers of dioecious species. 
ts =sterile, f= fertile, a = absent. 


*The placing of part of the Myrtales in the Rosiidae by Cronquist (1968) on the basis of the direction of stamen 
development may be found to be erroneous when other characters have been taken into consideration. Whether the 
androecium develops centripetally or centrifugally is a less fundamental character than Cronquist supposed (cf. 
Leins, 1964; Sattler, 1976; Tucker, 1976). 


100 N. K. B. ROBSON 


single vascular trace branching well above the torus. All ‘branching stamens’ that have been 
studied have proved to have separate staminal traces running from the region below toral 
level where the trunk trace dichotomises, e.g. Hibbertia (Dilleniaceae) (Wilson, 1965), 
Garcinia (Stebbins, 1974: fig. 12-1), and Vismia and Cratoxylum (Robson, 1956, 1972a). 
These stamen fascicle trunks are therefore the result of secondary fusion of filaments, as they 
are in Hypericum sects 25. Adenotrias, 28. Elodes and part of 30. Spachium. The tissue of 
the primary trunk remains at the base of some fascicles (e.g. in species of Camellia or 
Hypericum with distinct fascicles), but any higher fusion should be attributed to this 
secondary union. Thus, within-fascicle modifications have resulted in either (i) the complete 
elimination of the fascicle trunk and the merging of the constituent stamens to form a 
‘polyandrous’ androecium (e.g. in Hypericum sect. 20. Myriandra and most genera of the 
Bonnetioideae and Myrtaceae) or (ii) the secondary fusion of the filaments and sometimes 
the anthers to form a markedly fasciculate androecium (e.g. in some species of Hibbertia 
sect. Candollea (Wilson, 1965), Garcinia and the Moronobeoideae) or (iii) reduction of the 
fascicle to a single stamen (e.g. in Hypericum gentianoides, sect. 30. Spachium). 

As well as within-fascicle modification there is between-fascicle modification, in that the 
fascicles themselves may unite, either altogether to form a tube (as in Endodesmia, 
Calophylloideae, or Symphonia, Moronobeoideae) or in pairs 2+1+1+1 or2+2+41 (asin 
the Cratoxyleae and various sections of Hypericum). The latter modification has also 
occurred in the Cucurbitaceae (Violales), but after reduction of the fascicle to one stamen 
(Chakravarty, 1955). 

A summary of androecial variation in the Guttiferae is given in Table 3 (modified from 
Robson, 1972a). 


(c) Androecial trends 


The reasons for regarding the afasciculate androecium in Hypericum as derived were dis- 
cussed in Part I (Robson, 1977 : 303). It remains to outline the androecial trends in the genus 
as a whole (Fig. 20). 

The primitive condition of five free antipetalous fascicles (Fig. 20a) occurs constantly in 
the first seven sections except 2. Psorophytum, in which H. balearicum frequently has four 
fascicles due to the coalescence of the pair on either side of sepal 5 (Fig. 20b). The spatial and 
developmental reasons for this fusion are the same as those pertaining to the variation in 
sepal size and arrangement (p. 86) (see Hirmer, 1917; Breindl, 1934 : esp. fig. 33). The four- 
fascicled state seems to be relatively unstable, as it usually occurs only as an abnormality ora 
transitional evolutionary stage to the three-fascicled state (Fig. 20c). This is widespread in - 
the genus and results from the additional coalescence of the pair of fascicles on either side of 
sepal 4. The double fascicles have more stamens than the single one (unless numbers are very 
small), the one on the radius of sepal 4 being larger than that on radius 5 (Saunders, 1936). 
The three fascicles thus formed usually remain distinct and the configuration stable; but in 
two sections there has been a reversion from long-established trimery to pentamery, viz. sects 
13. Drosocarpium (the Cyprus population of H. repens = H. modestum Boiss.) and 
15. Thasia (H. thasium); and in sects 11. Campylopus (H. cerastoides) and 26. Humifu- 
soideum the number of androecial members is less stable. In H. cerastoides it varies from five 
to three, and the fascicles are frequently difficult to recognise owing to their uniting to form a 
very shallow ring (Fig. 20c, e-g). In 26. Humifusoideum, however, there is no ring- 
formation, and the individuality of the fascicles is lost merely by merging (Fig. 20f, g). As 
this alteration is less fundamental than the one to a distinct 2+2-+1 formation, it can be 
more easily reversed to a rather indefinite pentamery, a change that parallels a reversion 
from trimery to pentamery in the gynoecium: H. papuanum 3 (4-5), H. natalense 3-4-5, H. 
wilmsii 3-4, H. peplidifolium (4)5 (cf. Saunders, 1937; Robson, 1956, 19585, 1973a; Killick 
& Robson, 1976). 

A similar slight merging of the fascicles in sect. 30. Spachium has resulted in changes 
comparable with those in sect. Humifusoideum but more far-reaching. In the more primitive 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 101 


er = 
woe 


d 

Fig. 20 Androecial configuration trends in Hypericum: (a) 5 free fascicles; (b) 2+1+1+1 
fascicles, i.e. condensation along radius S5; (c) 2 +2 + 1 fascicles, i.e. condensation along radii $5 
and S4; (d) reappearance of 3 alternipetalous fasciclodes (sterile fascicles); (e) union of fascicles to 
form a narrow continuous ring; (f) merging of fascicles with partial obscuring of members (‘5 
obscure fascicles’); (g) merging of fascicles along with radial condensation (‘3 obscure fascicles’); 
(h) reduction of each fascicle to a single stamen; (i) union of fascicles with radial condensation to 
form a broad continuous ring; (j) elimination of one androecial member, resulting in tetramery. 


species, e.g. H. brasiliense, five distinct fascicles can sometimes be observed (Fig. 20a, f), 
and in two species that are rather isolated taxonomically (H. piriai, H. hilaireanum) there are 
three fascicles with the filaments in each one united for half their length or more (Fig. 20c). 
In most of the section, however, the stamens form a ring, initially continuous (Fig. 20e) but 
gradually becoming more irregular in the more advanced, smaller-flowered species (Fig. 20f, 
g). Eventually, in this evolutionary trend, the reduction in stamen number results in the 
reappearance of five fascicles (e.g. in H. oligandrum and H. pleiostylum); and these are 
ultimately each reduced to a single stamen in H. gentianoides, the type species of Sarothra 
L., which was placed by Linnaeus (1753) in his Pentandria Trigynia (Fig. 20h). In both sects 
20. Myriandra and 29. Brathys, on the other hand, the ring of stamens remains continuous. 
As will become evident later (p. 138), these sections show differing trends toward reduction 
in the size of the torus. In Brathys the condensation is mainly vertical and results in a 
relatively broad torus with a narrow ring of free stamens, and the perianth and androecium 


102 N. K. B. ROBSON 


remain constantly pentamerous (Fig. 20e). In Myriandra, however, the condensation is 
primarily lateral and produces a relatively narrow torus with a broad ring of free stamens 
(Fig. 201). There is therefore more radial congestion on the torus, so that tetramery becomes 
established in the three outer floral whorls (Fig. 20)). 

The reappearance in sects 25. Adenotrias and 28. Elodes of three sterile members of the 
outer whorl in connection with the evolution of specialised insect pollination (Fig. 20d) has 
been discussed already (Robson, 1972a) and will be mentioned again later (p. 123). 


(d) Form and glandularity 


The stamens of Hypericum and Santomasia, like those of most of the Bonnetioideae 
(Maguire, 1972) and of some of the rest of the Guttiferae (Calophylloideae, Vismieae, 
Cratoxyleae), have small, almost isodiametric tetrathecal anthers, which dehisce introrsely 
by longitudinal slits, and slender filaments. The stoutest filaments are found in H. bequaertii, 
but neither they nor the anthers vary in form in a taxonomically useful way. Unlike the 
essentially basifixed anthers of the Bonnetioideae, those of Hypericum are dorsifixed, the 
apparently basifixed ones in sect.4. Takasagoya resulting from heterogonic growth 
(Robson, 1973a). 

Terminating the connective is a flat or hemispherical gland which varies in colour from 
amber to red or black according to the amount of hypericin that it contains. Thus it is amber 
in sects 3. Ascyreia, 20. Myriandra and 30. Spachium, for example, and black in several 
sections including 9. Hypericum and 27. Adenosepalum. Comparable connective glands 
occur in some genera of the Bonnetioideae (Maguire, 1972: fig. 20), where they are 
cupuliform, and in the rest of the Hypericoideae, where they are sometimes paired (e.g. in 
Cratoxylum). In Neotatea neblinae (Bonnetioideae), Maguire (1972 : fig. 26) depicts what 
appears to be a glandular exudate at the apex of the anther. 


(e) Colour 


The stamens in Hypericum are nearly always the same colour as the petals, the exceptions 
being H. calycinum and its hybrid with H. patulum (H. x moseranum André) (sect. 3. 
Ascyreia) and H. pulchrum (sect. 18. Taeniocarpium), where the anthers are crimson. The 
orange colour in the anthers of H. x ‘Hidcote’, contrasting with the golden petals, is no doubt 
due to the influence of one of its parents, probably H. calycinum (see p. 170). 


(f) Length and movement 


The stamens are almost always shorter than the petals, exceeding them only in H. hircinum 
(sect. 5. Androsaemum), and their relative length is often a useful specific character. They 
vary in length within a fascicle, and so measurements in the specific descriptions will always 
refer to the longest stamen. 

The longer stamens are always on the outside of the fascicle or ring and are therefore the 
first to spread at anthesis. Later, when the stamens close up again, these longer ones may 
become entangled with the styles and effect self-pollination. 


(g) Staminodial members 


The occurrence (in Santomasia and other Hypericoid genera) or reappearance (in 
Hypericum sects 25 and 28) of conical or ligulate bodies between the stamen fascicles has 
been discussed above and in earlier works (Hochreutiner, 1918; Robson, 1972a, 1977a). 
Morphologically they are sterile antipetalous stamen fascicles, not stamens, and so the term 
‘staminode’ seems to describe them inappropriately. In previous publications (e.g. Robson, 
1967b, 1972a) I have used the word ‘fasciclode’ (i.e. a sterile fascicle), which describes the 
nature of the organ more accurately, if somewhat inelegantly. In Hypericum and some other 
genera they act as lodicules, expanding to spread the petals; but this word cannot be applied 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 103 


generally, and so ‘fasciclode’ is the better term to use, especially if the first ‘c’ is not 
pronounced. 


Ovary 
(a) Gynoecial elements 


The gynoecium in Hypericum, as in the rest of the Guttiferae, is superior and occupies the 
whole central part of the torus, i.e. there is no evidence of a disk, a perigynous zone or a 
gynophore. Like the members of the Bonnetioideae, Calophylloideae (except Mammea) and 
the other Hypericoideae, Hypericum has an ovary surmounted by elongate styles, each of 
which is terminated by a stigmatic surface. These styles vary in number from 5 to 2, as do the 
placentae, ovary lobes and capsule valves. They may be completely free or more or less 
coherent or united; but the stigmas remain distinct, or at least identifiable by the number of 
lobes, except in sect. 4. Takasagoya. There is therefore never any difficulty in determining 
the number of gynoecial elements (carpels, gynophylls, etc.) that are present in Hypericum. 
The variations in these gynoecial characters in the Guttiferae are shown in Table 4. The 
gynoecium of the Myrtaceae resembles that of the Hypericoideae and Bonnetioideae in 
length of styles and in placentation, differing essentially only in having an inferior ovary. 

The gynoecial elements will be termed ‘carpels’ in discussion for convenience, but without 
any implication of acceptance of the Candollean theory of the carpel. Indeed, gynoecial and 
fruit structure in Hypericum lends support rather to the Gonophyll Theory (Melville, 1962). 


Table 4 External characters of the gynoecium in the Guttiferae 


Taxon No. of elements Style length* Style and stigma 
uniont 
Bonnetioideae 5,3 ] flu 
Hypericoideae 
Hypericeae 5-2 l flu 
Vismieae 5 l f 
Cratoxyleae (5-4) 3 l f 
Calophylloideae 
Paramammea 4 l fu 
most genera (4) 2-1 l u 
Mammea (4) 2 S u 
Poeciloneuron 2, l f 
Moronobeoideae 5 ] u 
Clusioideae 
Clusieae (4) 5-6 (10) S,a ff 
Decaphalangium 5 Sp ff 
Allanblackia 5 S u 
Garcinieae (3) 4-5 (-12) s,a u 


*] = long (i.e. noticeably extended), s = short (not quite absent), a = absent 
f= free, u = united (when styles are absent this applies to stigmas) 
only male flowers described 


(b) Relationships with other whorls 


When one considers the exigencies of space on a developing floral receptacle, it is not 
surprising that a trend towards oligomery in Hypericum (Fig. 6) effects the innermost 
(gynoecial) floral whorl first. For example, in sect. 5. Androsaemum the outer whorls are 
5-merous and the gynoecium 3-merous. The statement by Keller (1925) that H. peplidi- 
folium (sect. 26. Humifusoideum) has 3 stamen fascicles and 5 styles, thus providing an 


1C4 N. K. B. ROBSON 


exception to this rule, was shown to be due to faulty observation (Saunders, 1937; Robson, 
1956). 

This meiomeric tendency is associated with the crowding along the radii of sepals 4 and 5 
noticed by Breindl (1934) and Saunders (1936), which was discussed in connection with the 
quincuncial development of the calyx (p. 86). Although there should theoretically be some- 
what greater crowding along radius 5 than along radius 4, the difference is slight. This 
explains why there is usually an evolutionary ‘jump’ from 5-mery to 3-mery, the inter- 
mediate 4-mery being an unstable state and never solely characteristic of a species (Fig. 21). 
Where it does occur (e.g. in sects 2. Psorophytum, 9. Hypericum and 26. Humifusoideum), it 
is associated with a 4-fascicled or irregular androecium (Fig. 6). The biological advantage of 
this arrangement, whereby the styles continue to alternate with the stamen fascicles or 
groups, should be noted. The extension of this meiomerous trend to 2-mery has occurred 
rarely in Hypericum, where it is associated with either 4-mery of the outer whorls (in 
sect. 20. Myriandra) or extreme reduction in flower size (in sect. 30. Spachium). 


K2 


Fig. 21 Transition from 5-mery (dotted lines) to 3-mery (solid lines) in the inner floral whorls of 
Hypericum (after Breindl, 1934). The crowding along sepal radii 4 and 5 is accompanied by a 
corresponding widening of the angle between sepals 1 and 3. Note that the theoretically equal 
angles are never found in practice, owing to the transition from opposite to quincuncial 
phyllotaxis (see p. 86). 


At this point it is perhaps desirable to warn against explaining these spatial relationships 
in causal terms. As Goebel (1898-1901 : 712, footnote) said [my translation]: ‘If we can see 
in a flower that more numerous primordia appear where there is more room at the floral 
apex, this does not necessarily mean that we can say that space relationships are responsible 
for the numerical relationships; we can just as well assume that there is more space provided 
where the floral apex is most disposed to form organ primordia. All ‘mechanical’ 
explanations are excluded in these relationships.’ 


: 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 105 


On the other hand, Goebel’s own suggestion that nutritional differences may be involved is 
equally unconvincing. The expression of a genotype can be very much modified by vari- 
ations in nutrition; but one cannot validly attribute differences between plants with different 
genotypes to such variations alone, although the genotypes probably influence the nutri- 
tional requirements of the respective plants. 


(c) Shape and proportions 


The ovary in Hypericum varies in absolute size and in shape from narrowly elliptic or 
narrowly ovoid to globose, but these variations are of only minor taxonomic value. Of 
greater value are the relative proportions of styles and ovary, which vary quite widely. In H. 
bequaertii and H. revolutum (sect. 1. Campylosporus), the styles are about 1°2-1-6 times as 
long as the ovary, whereas in H. monogynum (sect. 3. Ascyreia) they are 3°5—5 times as long 
as the ovary. This change in proportion is due more to a reduction in ovary length by half 
than to an increase in style length, although that, too, has occurred. In H. olympicum 
(sect. 10. Olympia) the relative style length (4-5 times) may be seen to be due almost entirely 
to a decrease in ovary size when compared with related species in sect. 3. Ascyreia. On the 
other hand, the styles of H. japonicum (sect. 30. Spachium), which are only 0:2-0°3 times as 
long as the ovary, appear to have attained their present proportions largely as a result of an 
absolute decrease in style length. 


(d) Union of styles 


The styles in the primitive Hypericeae are erect and appressed, but free (as in Santomasia) or 
only slightly coherent (as in Hypericum bequaertii), eventually turning outward near the 
apex. This state is comparable with that of primitive members of the Moronobeoideae 
(Pentadesma) and Calophylloideae (Mesua), where the union of the styles is more intimate, 
and with that of the Bonnetioideae, where the 5 styles of Ploiarium are free and outcurved 
but the 3-5 styles of the other genera are completely united. 

Within Hypericum there has been a trend towards fusion of the styles in two parts of the 
genus. In sect. 1. Campylosporus, complete union is achieved in H. synstylum, H. 
quartinianum and H. roeperanum; but the union is relatively loose, as the sections derived 
from this part of sect. | have styles that either are completely free and divergent from the 
base (sects 21. Webbia, 22. Arthrophyllum, 27. Adenosepalum, 28. Elodes) or separate as the 
fruit matures (sect. 20. Myriandra). In the other trend the fusion becomes more intimate, as 
the styles do not separate in fruit in H. monogynum, H. prattii or H. longistylum (sect. 3. 
Ascyreia), whilst in sect. 4. Takasagoya (which is derived from this group of species) the 
fusion is complete, even the stigmas having lost their individuality. From the same part of 
sect. 3 is derived sect. 7. Roscyna, in whch the fusion trend is reversed, even within one 
widespread variable species (H. ascyron), with the result that in the derivative sections 8. 
Bupleuroides, 9. Hypericum and 9a. Concinna (see p. 173) the styles are free and (in the last 
two) become gradually more divergent (Fig. 22). 


(e) Stigmas 


The stigmas in Hypericum bequaertii are subglobose, i.e. they are wider than the style. 
Modifications, which are often of taxonomic value, have led to extreme reduction in width 
(e.g. in sect. 20. Myriandra, where the stigma is narrower than the upper part of the style) or 
to elaboration in the form of capitate stigmas, which are characteristic of certain parts of 
sects 29. Brathys and 30. Spachium. In some species of these sections the broadening of the 
stigma is correlated with a gradual upward broadening of the style. 


(f) Glandularity 


The ovary walls of Hypericum include resin-containing canals or glands. Although they are 
often visible in the flower, they tend to enlarge and become prominent in fruit and will there- 
fore be discussed under that heading (p. 109). 


106 N. K. B. ROBSON 


CEy styles united in flower 


E 24 | a 14 | ae ic choe 


2 eel a rc 


Fig.22 Variation in stylar union in Hypericum. 


(g) Placentation 


The gynoecium in Santomasia and primitive species of Hypericum, as in the Bonnetioideae, 
consists of two parts: (i) a central columella comprising the five placentae and (11) five curved 
peripheral parts of the ovary wall, which are united to each other and to the placentae by 
radial septa and are prolonged to form the styles. These septa are thin near the junction with 
the placentae, at the point where they separate from them in fruit (Fig. 23a; Maguire, 1972: 
fig. 24j, k); and they are double structures, as is shown by their splitting longitudinally in 
fruit. In essence, therefore, the ovary comprises five curved valves (tegophylls—Melville, 
1962) surrounding and alternating with five boat-shaped placental structures (sagittate in T. 
S.), which bear numerous ovules on both sides distally and approximate to one another more 
or less closely. The placentae are united towards the base; but above they are merely 
mutually appressed and may separate again towards the apex (Fig. 24a). 

This state of placentation may be described as axile; but the loose association of the 
placentae makes a transition to true parietal placentation simple, and such a transition has 
occurred independently several times (Figs 23 h—k, 25). In other evolutionary lines the axile 
placentation has become more definite as a result of the more intimate union of the 
placentae (Figs 23c-g, 25). In some sections the placentation is wholly axile (e.g. in sects 17. 
Hirtella to 19. Coridium) (Figs 23c-e, 24a, b), in others it is constantly parietal (e.g in sects 
29. Brathys, 30. Spachium) (Figs 23i-k, 24d), whilst a third group shows transitional stages 
(e.g. sects 20. Myriandra, 26. Humifusoideum) (Figs 23b, h, 24c). In sect. 5. Androsaemum 
the placentation remains axile at the base but becomes parietal almost abruptly above. Other 
trends have resulted in the gradual change of the columella element from a boat-shaped 
structure with an ovuliferous margin (Fig. 23a), by elimination of the sterile part, to a rod 
(Fig. 23e, k) or even a partial rod (Fig. 238). 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 107 


Fig. 23 Trends in placentation in Hypericum (T.S. of ovary c. 1/3 above the base) (diagram- 
matic): (a) H. hookeranum; (b) H. olympicum,; (c) H. x desetangsii; (d) H. montbretit; (e) H. 
empetrifolium, (f) H. cerastoides; (g) H. pulchrum; (h) H. papuanum; (i) H. rigidum; (j) H. 
brasiliense; (k) H. elodes. 

The arrows indicate overall trends, not relationship between the species illustrated. 


Fig. 24 Ovaries of Hypericum species in L.S. (diagrammatic) (numbers indicate sections): (a) H. 
revolutum (1); (b) H. orientale (16); (c) H. prolificum (20); (d) H. elodes (28) (a x 5;b x 7°5;.c x 6°5; 
dx 8). 


108 N. K. B. ROBSON 


The placentation in the Vismieae and Cratoxyleae is basically similar to that described 
above. The fleshy fruit of the Vismieae is associated with thicker ovary walls and septa, 
whilst in the (capsular) Cratoxyleae the placentation is incompletely axile (Cratoxylum) or 
completely axile (Triadenum, Thornea) or axile with intrusive false septa (Eliea, cf. Baas, 
1970). The Myrtaceae, as has already been stated, have a basically similar ovary structure to 
that of the Bonnetioideae and Hypericum, with the difference that the ovary is sunk in the 
receptacle and hence inferior. The fundamentally different, apocarpous gynoecium of the 
Dilleniaceae, however, makes one wonder if Corner (1976) was not correct in denying that 
there is a close relationship between this family and the Guttiferae. 


= loosely axile 
ay werstrasans COMpletely axile 
------- + parietal 
ee 


Fig.25 Variation of placentation in Hypericum. 


(h) Ovules 


Each placenta margin in the primitive type of ovary is covered with numerous ovules, which 
vary in insertion from suberect near the apex to pendulous round the incurving base (Figs 
23a, 24a). As the placentae become more parietal, the sterile region between their ovule- 
ee margins diminishes and, when the placentation is wholly parietal, it disappears (Fig. 

Dupuy & Guédés (1975) have described supernumerary axial ovules in two members of 
sect. 3. Ascyreia, viz. H. calycinum and H. x ‘Hidcote’. These are four to six in number and 
occur on the top of the floral axis, i.e. within the placental region. They are vascularised from 
the stelar tissue remaining after the placental traces have departed. Dupuy & Guédes inter- 
pret these axial ovules as being foliar, not foliolar structures, i.e. they view them as 
representing an inner whorl of whole ‘carpels’, not extra ovules belonging to the placentae of 
the extant ‘carpels’. According to the Gonophyll Theory, however, they terminate branch- 
lets of a branch formed by a major dichotomy of the fertile branch of the gynophyll, the other 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 109 


major branch being the placental bundle, which vascularises the remaining ovules (see p. 
141). Axial ovules have not been reported in any other species of Hypericum. 

Another trend in Hypericum ovaries is towards a reduction in the number of ovules, a 
trend which roughly parallels the overall reduction in size. Extremes are reached in H. 
thymopsis (sect. 17. Hirtella) and H. russeggeri (sect. 25. Adenotrias), which have respec- 
tively 4-6 and 2 ovules per loculus. In the latter species one ovule is erect and the other 
pendulous (Jaubert & Spach, 1842-43 : t. 37). 

The ovules are anatropous and bitegmic, as in the rest of the Guttiferae (Corner, 1976; 
Philipson, 1974, 1977), and tenuinucellate, as in most of that family, with the micropyle 
formed by the exostome (Corner, 1976). 


Fruit 
(a) General 


In Santomasia and most Hypericum species the fruit is a somewhat woody or coriaceous 
septicidal capsule as in the Bonnetioideae but, unlike them, dehiscing at the apex, and 
usually with persistent valves. In four distantly related parts of Hypericum, however, a 
different type of fruit has evolved. In H. olivieri (sect. 17. Hirtella) the three ovary loculi are 
each 7-10-ovulate; but only one ovule usually matures into a seed, which is dispersed in the 
deciduous, subglobose, coccoid, capsule valve. Trends towards a baccate fruit have occurred 
thrice, but only one of them has resulted in the evolution of a true berry. In sect. 5. 
Androsaemum most of the species have a tardily and incompletely dehiscent capsule. H. 
androsaemum itself, however, has a fruit in which the pericarp is rather fleshy and colours 
during maturation to reddish-brown or black. H. x inodorum, of which it is a parent, shows 
similar but less extreme colour changes and is less succulent. Even in H. androsaemum, 
however, the ‘berry’ dries and can then be split into three valves with slight pressure from the 
fingers. The semi-fleshy fruits of H. reptans (sect. 3. Ascyreia) are even less baccate, but those 
of H. peplidifolium (sect. 26. Humifusoideum) are true berries. 

In the Vismieae, true berries are characteristic of Vismia (many-seeded) and Psoro- 
spermum (5-seeded), whilst Harungana has a fleshy drupe with five several-seeded pyrenes. 
The genera of the Cratoxyleae all have apically dehiscent septicidal capsules except Eliea, 
where the development of false septa has resulted in a capsule with irregular dehiscence, 
partly septicidal partly loculicidal (Gogelein, 1967; Baas, 1970). The valves are more or less 
persistent except in some species of Cratoxylum, where they are caducous. In the other tribes 
and subfamilies of the Guttiferae the fruit is mostly indehiscent, except in the Clusieae, 
where it is capsular and loculicidal, and Mesua ferrea L., one of the more primitive species in 
the Calophylloideae, where the woody capsule also splits loculicidally. 


(b) Styles 


In many species the styles persist on the capsule, then wither and may break off irregularly 
(e.g. Hypericum cerastoides, sect. 11. Campylopus). Others have a zone of weakness near the 
base of the style, above which it withers and usually breaks off cleanly (e.g. H. elongatum, 
sect. 17. Hirtella). In both cases the capsule length will be measured from the base to the 
point of withering. 


(c) Vittae and vesicles 


As was mentioned above (p. 105), the ovary wall contains resin channels (vittae) which in 
Santomasia and some sections of Hypericum remain inconspicuous (e.g. in sects 1. 
Campylosporus and 29. Brathys), i.e. the capsule is ‘not vittate’ (Fig. 26a). In others (e.g. 
sects 9. Hypericum pro parte and 17. Hirtella) they become prominent in fruit as vertical 
raised lines, sometimes with superimposed glands (Fig. 26b). The next stage in this 
evolutionary trend is for the lateral vittae to become diagonal while the dorsal ones remain 
vertical (e.g. in H. maculatum, sect. 9. Hypericum) (Fig. 26c). The vittae may then become 


110 N. K. B. ROBSON 


swollen (e.g. in H. perforatum, sect. 9. Hypericum or H. perfoliatum, sect. 13. Drosocarpium) 
(Fig. 26d) and break up into short streaks or dots (‘vesicles’) (e.g. in H. montbretii, sect. 13. 
Drosocarpium) (Fig. 26e). By comparison with its nearest relatives, H. sampsonii (sect. 9. 
Hypericum) seems to have evolved the vesiculate state directly from the vittate one; and in 
some species of sect. 30. Spachium (e.g. H. pauciflorum), one or two large vesicles develop on 
an otherwise smooth valve (Fig. 26g). Finally, although almost all these vittae and vesicles 
contain only amber resin, in H. richeri (sect. 13. Drosocarpium) some vesicles are blackish 
due to the presence of hypericin (Fig. 26f). 


Fig. 26 Capsules of Hypericum species, showing patterns of vittae and vesicles (see text): (a) H. 
revolutum (1); (b) H. elongatum (17); (c) H. maculatum (9); (d) H. perfoliatum (13); (e) H. 
montbretii (13); (f) H. richeri (13); (g) H. diffusum (30) (all x 4). 

Seeds 
(a) Number 


The number of seeds in a Hypericum capsule is sometimes very large (often well over 1000, 
cf. Salisbury, 1963); but trends in various groups have resulted in a reduction in number to 
the extremes found in H. russeggeri (sect. 25. Adenotrias) and H. olivieri (sect. 17. Hirtella), 
where only 6 and 3 seeds, respectively, are produced in each capsule. 


(b) Shape, size and colour 
The seeds in the Hypericoideae and Bonnetioideae, unlike those of the other Guttiferous 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 111 


subfamilies, are small and narrowly cylindric to ovoid-cylindric or ellipsoid. The longer ones 
may be slightly curved, especially if they are borne towards the base of the placenta; but this 
variation has no great taxonomic significance, despite the name of Hypericum sect. | 
(Campylosporus). In Hypericum they range from 1:5 mm long in H. bequaertii (sect. 1. 
Campylosporus) to 0°3 mm long in H. gentianoides (sect. 30. Spachium). They vary in 
colour from yellow-brown to red-brown or dark purplish brown. 


(c) Appendages 


Winged seeds occur in most Bonnetioid and Hypericoid genera with dehiscent fruit, usually 
in the more primitive members (e.g. in Cratoxylum but not in Triadenum or Thornea). 
Their apparently primitive status would seem to give support to Corner’s (1976) theory that 
the wing is a vestigial arillar structure (true arils occur in the Clusieae). In some species (e.g. 
Cratoxylum arborescens (Vahl) Blume or Mahurea casiquiarensis Spruce) the wing 
surrounds the rest of the seed completely or almost so, in others (e.g. Cratoxylum maingayi 
Dyer) it is confined to one side, whilst in still others (e.g. Ploiarium alternifolium (Vahl) 
Melchior) it consists of terminal prolongations joined by little more than a carina. The two 
last-mentioned stages are found in Hypericum*, but only in the more primitive sections (Fig. 
27), the wing being thin and papery and thus unlike the cartilaginous wings with a peripheral 
vein that occur in Cratoxylum (Plate Ic). 

Even if the wing has been reduced to a carina or less, the seed may have apical and some- 
times basal prolongations (e.g. in H. geminiflorum, sect. 4. Takasagoya) or apiculi (cf. Plate 
1b). In sect. 25. Adenotrias the basal apiculate expansion has evolved into a fleshy whitish 


GEASS winged (terminal) 
, carunculate 
Ka | > carinate (lateral) 


Fig. 27 Variation in seed appendages in Hypericum. 


*] have seen no seed of Hypericum with a complete marginal wing such as is illustrated by Keller (1925 : fig. 73P). 


112 N. K. B. ROBSON 


caruncle (Plate 2c). Although this change to specialised seed dispersal occurs in the very 
section that has evolved heterostyly, there is no obvious relationship between the two 
specialisations. No other Hypericum has a specialised means of seed dispersal (as opposed to 
fruit dispersal). 


(d) Testa sculpturing 


The seeds of Hypericum have the thick-walled stellate tegmen cells characteristic of all the 
Hypericoideae (except Psorospermum, where they are fleshy), as well as the Clusieae and 
probably the Bonnetioideae (Corner, 1976). The testa, however, is the part of the outer 
integument that provides characters that are infragenerically useful. Its inner layer is 
inconspicuous; but the cells of the outer epidermal layer have brownish to blackish tannin 
contents, and the radial and inner walls are more or less thickened (Ohlendorf, 1907). As the 
testa dries and matures, these thickened walls resist collapse, forming a reticulum of 
depressions in the testa; and the reticulate pattern so formed is sometimes characteristic of 
whole sections. 

In the primitive state (e.g. in Santomasia and Hypericum sect. 1. Campylosporus) the 
exotestal cells form roughly defined lines and have relatively thin walls (i.e. the testa pattern 
is ‘linear-reticulate’) (Fig. 28 RE; Plate la—c). Three types of modification of this pattern 
can be observed: 

(i) deformation of the lines of cells, making the testa merely ‘reticulate’, not ‘linear- 
reticulate’ (not differentiated in Fig. 28) (Plates 1d, 2b); 

(i1) thickening of the lateral cell walls, often accompanied or preceded by lateral elonga- 

tion of the cells (Plates le, 2d); 

(iii) protrusion of the outer cell walls (Plate 2a). 

Where thickening is confined to the longitudinal walls, the pattern becomes ‘scalariform- 
reticulate’, as for example in some species of sect. 30. Spachium (Rodriguez-Jiménez, 
1974 : plate 2b, c) (Fig. 28S—-RE; Plate 2e). Where this longitudinal thickening has become 
more pronounced, so that the seed looks ridged, the pattern is termed ‘ribbed-scalariform’ 
(Rodriguez-Jiménez, 1974 : plate 2a) (Fig. 28RI-S; Plate 2d). 

Where all the walls are thickened, so that the depression between them is round, the testa 
is ‘foveolate’ or ‘linear-foveolate’, a common state in Hypericum (Fig. 28FO; Plate 1f). 
Where the outer cell walls of a foveolate testa have begun to protrude, a ‘rugulose’ pattern is 
produced (in sects 12, 18-19, 22) (Fig. 28RU; Plate 2c); and more pronounced protrusion 
results in a ‘papillose’ testa (in sects 17-19) (Fig. 28PA; Plate 2a). The apparent absence of 
rugulose seeds from sect. 17. Hirtella, which in other characters is basic to sects 18-19, 
suggests that they probably occurred in the basic stock from which these three sections 
evolved, but that the trend to papillose seeds was established earlier in sect. 17. Hirtella than 
it was 1n the other two sections. 


(e) Embryo 


Although seeds of the Guttiferae are usually said to be exalbuminous, Corner (1976) points 
out that a layer 1-2 cells thick of nuclear oily endosperm is present in the Hypericoideae and 
sometimes in the other subfamilies, and that it may not be completely absent in Ploiarium 
(Bonnetioideae). Prakash & Lau (1976), however, make no mention of it in their paper on P. 
alternifolium. 

The embryo in Hypericum, as in the Bonnetioideae and other Hypericoideae except 
Psorospermum (Vismieae), is slender and straight with equal, plano-convex cotyledons 
which are somewhat shorter than the hypocotyl. In Psorospermum the seeds are reduced in 
number to one in each loculus of the 5-locular baccate fruit and are much larger than in the 
ancestral Vismia; and in the embryo the cotyledons vary in the three subgenera from equal, 
symmetric and accumbent (subgen. Psorospermum) to either unequal, asymmetric and 
incumbent (subgen. Parasorospermum) or equal and inrolled to conduplicate (subgen. 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 


Plate 1 Seeds of Hypericum, 1 (numbers indicate sections): (a) H. revolutum (1); (b) H. 
acmosepalum (3); (c) H. ascyron (7); (d) H. maculatum (9); (e) H. richeri (13); (f) H. linarifolium 
(14) (a x 46, b x 53,c x 51,dx 55,ex 46, fx 90). 


N. K. B. ROBSON 


Plate 2 Seeds of Hypericum, II (numbers indicate sections): (a) H. hirsutum (18); (b) H. 
hypericoides (20); (c) H. aegypticum (25); (d) H. elodes (28); (e) H. moranense (30); (f) H. 
mutilum (30) (a x 60, b x 63, c x 40,d x 75,e x 130, fx 110). 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 115 


Fig. 28 Variation in seed testa in Hypericum: FO foveolate, PA papillose, RE reticulate, RI-S 
ribbed-scalariform, RU rugulose, S-RE scalariform-reticulate. 


Afropsorospermum) (Bamps, 1966). According to Vestal (1938), the cotyledons are longer 
than the hypocotyl in the Vismieae and Cratoxyleae, but shorter than it in the Hypericeae. In 
the other subfamilies of the Guttiferae the embryo structure varies widely and becomes 
highly specialised (cf. Brandza, 1908). 


4. Pollen morphology 
By G. C. S. Clarke 


Introduction 


This is a summary of a more extensive paper on the pollen morphology of the genus 
Hypericum which will, it is hoped, appear in this journal in due course. To avoid repetition, 
most of the details and the illustrations have been omitted from the present account; they 
will be found in the main paper. 

The pollen grains described here have been studied by means of the light microscope and 
the scanning electron microscope, after preparation by the standard acetolysis technique as 
refined by Reitsma (1969). The pollen samples have all been taken from herbarium speci- 


mens which will be listed in the main account. About 250 species have been examined. 


General description 


The standard pollen type throughout the genus is tricolporate with a microreticulate or 
reticulate pattern of ornamentation. Pollen of this kind is common throughout the dicotyle- 
donous angiosperms, so Hypericum pollen is not particularly distinctive and there is little 


116 N. K. B. ROBSON 


evidence from pollen morphology for the discreteness of Hypericum as a genus. Some genera 
of the Guttiferae are more distinct in their pollen morphology (Clarke, 1975); a few have 
porate rather than colporate grains and a few are characterised by having more than three 
apertures. 

The only divergence within the genus Hypericum from the standard tricolporate plan is 
the tendency for some species to produce pollen grains with more than three apertures, and 
in which the strict polarity of the standard grains has disappeared. There is circumstantial 
evidence that this phenomenon is linked with cytological factors that influence meiosis 
(Clarke, 1975). In many of the species where they occur, the irregular grains form only a 
small proportion of the total, but in a few species practically all the grains are of the 
irregular type. This is discussed below (see pollen type XI). 

A generalised description of the features common to the pollen grains of all the Hypericum 

species I have seen is as follows (terminology according to Erdtman, 1971; Faegri & Iversen, 
1975; Reitsma, 1970): 
Pollen tricolporate, spheroidal to prolate. Outline in polar view more or less circular to 
three-lobed with the apertures set between the lobes, or triangular; in equatorial view 
elliptic, rectangular-elliptic or rhombic. Ectoapertures colpi, long to very long, parallel- 
sided or, more often, widest at the equator; often partly covered at the equator by sexine 
extensions; margin sometimes thickened near equator; colpus membrane smooth or 
granular. Endoaperture a lalongate colpus, porus or lolongate porus, sometimes with short 
lateral extensions; margins often thickened across colpus membrane. Exine usually rather 
thin, nexine about as thick as sexine; columellae short, unbranched. Ornamentation micro- 
reticulate, reticulate or a tectum perforatum; muri often broader at the base than above, 
simpli- or duplicolumellate; lumina irregular in outline, sometimes enclosing free 
columellae. Size—Polar axis 13-45 um; equatorial diameter 7-28 um. 


Variations in pollen morphology 


Within the basic morphological scheme outlined above there is sufficient variety for eleven 
pollen types to be described. The types have been numbered from I to XI but there is no 
special significance in the numbering and it is not intended to imply a morphological or 
evolutionary sequence. The diagnostic characters of the types are as follows: 


Pollen type I: Grains prolate-spheroidal to subprolate. Endoaperture a porus with very small lateral 
extensions. Ornamentation a tectum perforatum or microreticulum; tectal perforations regularly 
spaced. 


Pollen type II: Grains prolate-spheroidal to subprolate. Endoaperture a porus, often more or less 
lalongate, with very small lateral extensions. Ornamentation a tectum perforatum or microreticulum; 
tectal perforations grouped together. 


Pollen type III: Grains subprolate. Endoaperture a lalongate colpus. Ornamentation a tectum 
perforatum or microreticulum; tectal perforations grouped together. 


Pollen type IV: Grains subprolate. Endoaperture a lalongate colpus. Ornamentation a tectum 
perforatum or microreticulum; tectal perforations regularly spaced. 


Pollen type V: Grains spheroidal or prolate-spheroidal. Outline in polar view triangular with concave 
sides. Endoaperture a lalongate colpus. Ornamentation a tectum perforatum or microreticulum; tectal 
perforations regularly spaced. 


Pollen type VI: Grains perprolate or prolate. Endoaperture a large porus, more or less circular or 
lolongate. Ornamentation microreticulate or reticulate; lumina regularly spaced. 


Pollen type VII: Grains very small, prolate. Endoaperture a large lolongate porus. Ornamentation 
microreticulate or reticulate; lumina regularly spaced. 


Pollen type VIII: Grains prolate. Endoaperture a very large lolongate porus or colpus. Ornamentation 
reticulate and microreticulate; lumina regularly spaced. 


: 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 7 


Pollen type IX: Grains very large, subprolate or prolate. Endoaperture a more or less lalongate porus 
with short lateral and meridional extensions. Ornamentation reticulate; lumina regularly spaced. 


Pollen type X: Grains prolate or subprolate. Endoaperture a more or less lalongate porus with short 
lateral and meridional extensions. Ornamentation microreticulate; lumina regularly spaced. 


Pollen type XI: Grains all of irregular form and variable shape. Apertures varying in number from 2 to 
12; arranged in many different ways. 


Discreteness of the pollen types 


Some of the eleven pollen types are very distinct and there is no difficulty in separating them 
from all the others. Examples of this are types V, VI, VII, VIII and IX. Some types are 
distinct in the majority of cases, but exceptional specimens may link them with other types. 
Thus type II is normally very distinct, but a few species have pollen grains which are 
morphologically intermediate between types II and I. Similarly, some species are inter- 
mediate between types III and IV. Types I, IV and X can also be hard to separate because the 
structure of the endoaperture is difficult to observe. 

Type XI is anomalous since it is defined on different criteria from the other types. Its 
recognition depends on the presence of an overwhelming proportion of irregular pollen 
grains in a specimen. In effect, what this means is that when a species does not produce 
normal pollen grains it cannot be categorised in the same way as other species. Some species 
of other types, notably II and X, produce a small proportion of irregular grains, but the 
remaining grains can be assigned to a type in the normal way (Clarke, 1975). 


Distribution of the pollen types amongst the sections 


The level at which pollen morphology relates to taxonomy in the genus Hypericum corre- 
sponds most closely to the section. Table 5 and Fig. 29 show how the pollen types are 
distributed amongst the 31 sections of the genus recognised by Robson. The species of some 
sections all have pollen of a single type (sects 20. Myriandra, 28. Elodes and 25. Adenotrias 
are examples of this), while those of other sections, notably sect. 3. Ascyreia, are divided 
between several types. In the same way, some of the pollen types, such as II or V, are found 
exclusively in species from a single section, while others, such as III, IV or particularly X, are 


found in a number of sections. In one or two cases the majority of the species in a section fall 


into one pollen type, but one or two species are anomalous and have pollen of another type. 
Sect. 9. Hypericum is a case in point: all the species I have examined fall into type X with the 


exception of H. epigeium, which has pollen of type IV. This kind of distribution pattern 


remains unexplained. 


Relationships between the pollen types 


The morphology of the eleven pollen types suggests that they can be grouped together in a 
| Way that may reflect relationships between them. Four main groups can be formed. The first 
| group includes pollen types I and II, which have a similar form of endoaperture. The second 
brings together types III, [IV and V, which are again linked by the form of their endo- 


apertures. The third includes types X, IX and XI, of which IX and XI seem to be special- 
| lsations derived from X. The fourth group combines types VI, VII and VIII; types VII and 


VIII, with their large endoapertures and characteristic ornamentation, seem more closely 
linked to each other than either is with type VI, which shares some features with them and 
other features with type I. 


118 


N. K. B. ROBSON 


Table5 Distribution of the pollen types amongst the sections of Hypericum 


Sections 


Pollen types 
II Ill IV Vv VI Vil VUI Ix 


Campylosporus 
Psorophytum 
Ascyreia 
Takasagoya 
Androsaemum 
Inodorum 
Roscyna 
Bupleuroides 
Hypericum 
Concinna 
Olympia 
Campylopus 
Origanifolia 
Drosocarpium 
Oligostema 
Thasia 
Crossophyllum 
Hirtella 
Taeniocarpium 
Coridium 
Myriandra 
Webbia 
Arthrophyllum 
Triadenioides 
Heterophyllum 
Adenotrias 
Humifusoideum 
Adenosepalum 
Elodes 

Brathys 
Spachium 


The distribution of morphological characters within these four groups suggests that there 
may be a sequence from the first group, the most primitive, to the fourth which is the most 
advanced. The arguments in favour of this conclusion will be developed in the main paper. 


Xx 


t++eetet 
+ 


Taxonomic implications of the pollen morphology 


Although every species in the genus has pollen which is constructed on a basically similar 
pattern, the variations described here suggest a number of taxonomic conclusions. Only the 


most general of these will be summarised here. 


1. Only in exceptional cases does pollen morphology contradict the way the genus has been 


subdivided into sections by Robson. 
2. Sect. 1. Campylosporus seems to contain the species with the most primitive pollen. 
3 


. Sects 29. Brathys and 30. Spachium contain the species with the most advanced pollen. 
They have a good deal in common with sect. 20. Myriandra, but with few others. 
4. Sect. 3. Ascyreia is morphologically rather variable, since it brings together species with 


four different pollen types. 


5. There are a large number of sections which are apparently closely related since all the 


species they include have pollen of type X. 


XI 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 119 


Fig. 29 Distribution of pollen types in Hypericum (see text). 


5. Biology of flower and fruit 
Pollination 


(a) Unspecialised flower 
The flowers of Hypericum are probably all nectarless and open to visitation by any insects. 


_ They are therefore typical ‘pollen flowers’ with relatively numerous stamens. The petals 


have no honey guides or strong odour that might attract insects, although the whole plant 
sometimes gives off an aromatic scent in hot weather. This is frequently reminiscent of curry 


_ (e.g. the South African name for H. revolutum (sect. 1. Campylosporus) is Curry Bush— 


Killick & Robson, 1976); but it may be the less pleasant ‘he-goat’ aroma of H. hircinum 
(sect. 5. Androsaemum). The latter is due to caproic acid (n-hexanoic acid) or caprylic acid 


_ (n-octanoic acid). For these reasons the flowers of Hypericum are relatively primitive; and 


further primitive characters are (i) the prevalent yellow colour of the petals and inner organs 
and (11) the homogamy and apparent self-compatibility of most species. 

The flowers of species such as H. kouytchense (sect. 3. Ascyreia) are of the type described 
by Faegri & van der Pijl (1971) as primitive brush blossoms (Fig. 30a). In such flowers the 
numerous stamens provide a platform over which the insect visitor moves, the anthers 
brushing against the underside of the insect’s body and transferring pollen to it. This pollen is 
then deposited by the insect on an adjacent stigma or on one in another flower. 

The ‘platform’ is produced by the gradual inward movement of the stamens as the anthers 
dehisce, starting with the innermost ones, which in nearly all species are the shortest. As 
these do not attain the level of the stigmas, cross-pollination is favoured. In H. calycinum 
and some other Hypericum species with relatively primitive flowers, the differentiation of 


120 N. K. B. ROBSON 


height is not great; but in the more advanced flower of H. perforatum, for example, it is 
considerable (Fig. 30b). This relationship between stamen length and stigma level (i.e. style + 
ovary length) appears to exist in most open-pollinated species of Hypericum. In some ‘wide’ 
hybrids (e.g. H. x moseranum Luquet ex André and H. ‘Hidcote’, both sect. 3. Ascyreia), 
however, the anthers do not ever reach the level of the stigmas, which indicates that the 
mating of two species with distinct floral structures has resulted in biological disfunction 
(quite apart from any genetical or chromosomal disfunction that has occurred). A second 
exception is found in at least some species of sect. 29. Brathys (e.g. H. laricifolium), in which 
the stamens fall far short of the stigmatic level. From the observation that the filaments in 
such flowers are rather thick and crumpled, however, it would seem that they have under- 
gone some sort of contraction, possibly associated with self-pollination at anthesis. 


Fig. 30 Pollination of unspecialised Hypericum flowers (one sepal and two petals removed): (a) 
H. calycinum (x 2:5); (b) H. perforatum (x 4). 


As the outer stamens move inward, the anthers come into contact with the stigmas, thus 
ensuring self-pollination if cross-pollination has not occurred (Fig. 30b). In some species 
(e.g. H. linarifolium and H. humifusum, sect. 14. Oligostema), unfavourable weather 
conditions result in pseudocleistogamy, i.e. the flowers do not open and are automatically 
self-pollinated (Ivimey-Cook, 1963). The styles in bud are erect and in contact with the 
anthers. In primitive species they remain erect or more-or-less united, so that no movement 
is necessary for self-pollination to occur, other than a slight outcurving at the tip. In several 
sections (e.g. 10. Olympia) the petals close up as well as the stamens, resulting in a second 
bud-like stage of anthesis. This movement, however, does not appear to be necessary to 
ensure self-pollination (although it may help to produce it), as the species in which the petals 
and/or stamens are deciduous (e.g. in sect. 3. Ascyreia) are not noticeably less fertile, under 
similar conditions, than those with a second bud-stage. 


(b) Insect visitors 


Such flowers as those that have just been described are typically visited by the less- 
specialised insects, of which the Syrphidae (Diptera) are the commonest visitors to 
Hypericum. They are particularly associated with yellow flowers (Faegri & van der Pijl, 
1971), and the well-known irregularity of their visits may have been a factor in the evolution 
(? or retention) of a self-compatible breeding system. Bombylids, which are also frequent 
visitors, are more specialised; but they commonly visit primitive flowers for pollen. Miiller 
(1883), moreover, observing that these flies frequently put their tongues into the flowers of 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 


Plate 3H. ‘Hidcote’, photographed in ultraviolet light to show the ‘bull’s-eye’ (bee-yellow) at the 
centre of the flower and the contrasting outer (bee-purple) region. 


122 N. K. B. ROBSON 


H. perforatum, suggested that they were boring into the soft tissues. They would thus obtain 
sap but not, of course, nectar. 

For lists of specific insect visitors, the reader is referred to Knuth (1908) and Faegri & van 
der Pil (1971). 

The aromatic scent characteristic of some species of Hypericum may, as has been stated, 
attract insect visitors to the vicinity of the plant; but the attraction of the flowers themselves 
is probably wholly visual. There is no internal contrast to the yellow colour, such as would 
be provided by e.g. honey guides; but the anthers are sometimes darker yellow to orange (e.g. 
in H. hookeranum, sect. 3. Ascyreia) and rarely may be reddish (H. calycinum, sect. 3. 
Ascyreia and H. pulchrum, sect. 18. Taeniocarpium). The colour contrasts are on the outside 
of the petals, in the form of red tinges or veining, which are invisible when the flower has 
opened. In bud, however, they are conspicuous. As they are confined to that part of each 
petal that is exposed, the whole bud appears dark red or red-veined. Eisner et al. (1973) have 
shown that these red patterns appear dark in ultraviolet light, in contrast to the rest of the 
outside of the petal. The centre of the open flower (petal bases, stamens, ovary) is also 
ultraviolet-dark, although it is not differently coloured to the human eye in normal light. To 
the visiting insect, however, the open flower is a ‘target’, with a yellow ‘bull’s-eye’ 
surrounded by ‘bee-purple’ (Plate 3). The unopened buds, in contrast, are wholly yellow and 
therefore less attractive from a distance. As these also differ in form from the open flowers, 
insects would have little difficulty in distinguishing and avoiding them. 


(c) Floral specialisation 


In the species with large flowers the single flower is the pollination entity (Faegri & van der 
Pijl, 1971); but, in more advanced species, reduction in size is accompanied by various 
degrees of aggregation. As there is a parallel reduction in the number of stamens in each 
flower, this floral aggregation will tend to retain an efficient ‘brush’ pollination mechanism. 
The most extreme aggregation is found in two species of sect. 30. Spachium from Uruguay 
and adjacent Brazil and Argentina, H. myrianthum and H. tamariscinum Cham. & Schlecht. 
(= H. notiale, H. pelleterianum), where the very small flowers are in a dense flat corymb. H. 
scabrum (sect. 17. Hirtella) has dense flat corymbs of larger flowers, whilst in the closely 
related H. capitatum the dense inflorescence is broadly pyramidal to capitate. 

Another specialisation concerns the styles, which, as we have seen, are erect with out- 
curved tips in the primitive species. In all evolutionary lines there is a tendency for the styles 
to spread more widely, and this is most noticeable in the sections (e.g. 9. Hypericum) where 
there are ‘three’ stamen fascicles. Here the three styles diverge from the base between the 
fascicles, favouring cross-pollination, only later assuming a more erect position that allows 
the longer (outer) stamens to close up and effect self-pollination. In sections where the 
stamens are irregularly distributed (e.g. 30. Spachium), the relation between the positions of 
style and stamens is less close, as it also is in those sections with transitional floral structure 
(e.g. 5. Androsaemum: 5 fascicles, 3 erect styles; 8. Bupleuroides: ‘3’ fascicles, 3 erect styles). 
In sect. 20. Myriandra, where the fascicles have merged completely, the styles are com- 
pletely appressed except in the more advanced species with a tetramerous perianth, where 
they eventually curve at the apex. 

In the two sections where specialised pollination occurs (sects 25. Adenotrias and 
28. Elodes), the character syndrome in the flowers of each section is remarkably similar 
(Robson, 1972a). In both, the sepals remain stiff and erect, so that the petals can spread out 
distally only. The corolla is thus pseudotubular; and, correlated with this development of a 
‘corolla tube’, the filaments in each stamen fascicle have become united to above the middle 
(Fig. 31). The petals have evolved a ligulate outgrowth (entire in Adenotrias, trifid in Elodes) 
which has juicy tissue in its axil and may be a source of nectar, although I have not observed 
free nectar in either H. aegypticum or H. elodes. In H. aegypticum, however, the ligule is 
appressed to the ovary by the development of a wedge of connecting tissue. The space 
between adjacent wedges forms a groove that would guide a probing insect tongue towards 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 123 


Fig. 31 Pollination of specialised Hypericum flowers (one sepal and two petals removed): (a) H. 
aegypticum (short-styled); (b) H. aegypticum (long-styled); (c) H. elodes (homostyled) (all x 4). 


the base of the ovary, where it would be deflected laterally to a space between two of the 
three hypogynous ‘scales’. These are entire in Adenotrias, bilobed in Elodes; and, whatever 
their morphological status may be (see p. 102 and Robson, 1972a), these act as lodicules, by 
swelling during development, thus helping to expand the flower (Hochreutiner, 1918). The 
three styles diverge so that the stigmas are between the ‘three’ fascicles and would be 
encountered by a pollen-laden probing insect. 

H. aegypticum and the other species in sect. Adenotrias are more highly evolved than H. 
elodes in exhibiting diheterostyly (Fig. 31a, b), combined, at least in H. aegypticum, with a 
moderately strong incompatibility system (Ornduff, 1975). Ornduff has shown that there is a 
good reciprocal correspondence between the positions of anthers and stigmas of the ‘pin’ and 
‘thrum’ forms of the flower, and that the pollen grains too are dimorphic. Those from anthers 
of ‘thrum’ flowers are on the average considerably larger than those from ‘pin’ flowers. In 
addition, ‘pin’ flowers produce an average of 1°7 times as many pollen grains as do ‘thrum’ 
flowers. Self-pollinations and own-form-pollinations produced 31°6-47°5% fruit-set and an 
average of 2:7-5°8 seeds per pollination, whereas the corresponding figures for other-form- 
pollinations were 83:3-100% and 23-9-27:0 respectively. 


Breeding systems 


Relatively little information is available on breeding systems in Hypericum. Self- 
compatibility is apparently widespread in the genus but not universal. The moderately 
strong incompatibility system in H. aegypticum has just been discussed; and Salisbury (1963) 
has discovered one in H. calycinum (sect. 3. Ascyreia). The latter, he suggested, is due to (i) 
degrees of failure in embryo development and (ii) inefficient endospermic nutrition during 
the attainment of the resting stage. Myers (1963), however, found H. calycinum and various 
other species and hybrids of Hypericum and Triadenum all to be pollen-tube/style 
compatible; and Culwell (1970), while reporting that the eastern North American species of 
sect. 9. Hypericum were all self-fertile, suspected that at least some of them might be pseudo- 
gamous (see p. 172). 

The flowers of Hypericum, like those of other genera of the Hypericoideae, are bisexual, 
and specialisation in this subfamily tends towards heterostyly, not dioecy. The occurrence of 
pistillody of the stamens in H. nudiflorum (sect. 20. Myriandra) reported by Rehder (1911) is 
no doubt, therefore, due to an isolated genetical aberration and is not an indication of an 
evolutionary trend. For other breeding aberrations, see 6. Development (p. 125). 


124 N. K. B. ROBSON 
Seed dispersal 
(a) Gravity 


Seeds of Hypericum species are cylindric to ellipsoid, small (about 0-3-1:5 mm long) and 
light (those of native and naturalised British species range from 0:00002 to 000058 g on 
average—Salisbury, 1942, 1963); and they are usually shed from the plant by the septicidal 
dehiscence of a capsule. Although they are small and light, they are not sufficiently so to be 
dispersed for long distances by wind without morphological adaptation—which most of 
them do not appear to have. The majority of seeds, therefore, fall near the parent plant. 


(b) Wind 


Although most seeds have no morphological adaptation for wind-dispersal, those of the 
more primitive species have a prominent carina or a narrow unilateral membranous wing or 
terminal appendage (e.g. in sect. 5. Androsaemum). No observations have been published 
that show these to be effective in wind-dispersal, but they may be expected to assist it to a 
limited extent. On the other hand, Ridley (1930) cites records of H. hirsutum (sect. 19. 
Taeniocarpium) and H. perforatum (sect. 9. Hypericum) growing on walls at about 6 m 
above the ground, well above the raindrop splash-zone; and so even wingless seeds may be 
carried for short distances by wind; Salisbury (1952) states without further comment that the 
(unwinged) seeds of H. montanum (sect. 27. Adenosepalum) are dispersed by wind, as does 
Ivimey-Cook (1963) for H. linarifolium (sect. 14. Oligostema). 


(c) Water 


H. elodes (sect. 28. Elodes) grows in damp mud or water and can flower where the water is 
not too deep (Gliick, 1911). In shallow water, when the capsules dehisce in August, the seeds 
float on the surface for about three days and can be dispersed by waves (Ohlendorf, 1907). 
Subsequently they sink and overwinter in the mud at the bottom, the sclerotic layer of the 
testa preventing their rotting. 


(d) Animals (external) 


(i) Birds. Many species of Hypericum besides H. elodes grow in damp mud, and their seeds 
might therefore be expected to adhere to the feet or feathers of wading birds and waterfowl. 
H. elodes has a limited seed-production, so that its occurrence in ponds that become dry is 
probably due to vegetative dispersal by ducks (Ridley, 1930). Other species of damp habitats 
are mostly herbs belonging to sect. 30. Spachium (e.g. H. canadense), and it is significant 
that this section is the only one in the genus to show evidence of long-distance dispersal. In 
contrast to H. elodes, such plants of lake- or riverside or marsh are probably propagated 
wholly by seed. Several North American members of this section have been found in 
scattered localities in Europe and western Asia (Caucasia) in the last 150 years or so 
(Gorshkova, 1949; Heine, 1962), and, although relict status (i.e. perglacial survival) has been 
proposed as the most likely explanation of the occurrence of H. canadense in Ireland (Webb, 
1958), the generally haphazard European distribution of these species suggests that carriage 
of seed by wading birds or waterfowl is the most likely explanation of their isolated occur- 
rences. The existence of wide disjunctions elsewhere in the section (see p. 206 and Robson, 
1977a) provides supporting evidence for this hypothesis. Of course, accidental human intro- 
duction cannot be ruled out in some cases (cf. Webb & Halliday, 1973). There is, however, a 
complete absence of direct evidence of bird-transport (no seeds of these species have been 
found on migratory birds, for example), but the circumstantial evidence is strong. 


(ii) Mammals. The absence from Hypericum seeds of adaptations favouring adherence to 
wool or hair indicates that mammals are not likely to be effective dispersal agents. There is 
one record of H. triquetrifolium (sect. 9. Hypericum) growing on wool-heaps in Montpellier 
(Thellung, 1912); but it is more likely, with such a divaricately branched species, that the 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 125 


whole fruiting plant (or part of it) adhered to the wool than that the seeds alone were 
transported. 


(iii) Insects. In the seeds of Hypericum aegypticum and the other species in sect. 25. 
Adenotrias, the terminal appendage has become a hard white caruncle. No doubt this is 
attractive to ants, which are therefore likely to aid the dispersal of these species. 


(e) Animals (internal) 


In three sections of Hypericum, as was mentioned above (p. 109), the capsule has evolved a 
fleshy wall and become tardily dehiscent or indehiscent. There are no recorded field observ- 
ations on H. reptans (sect. 3. Ascyreia) or H. peplidifolium (sect. 26. Humifusoideum) to 
show which animals utilise these berries as food, and neither species has relatives with tardily 
dehiscent fruit. 

In sect. 5. Androsaemum, however, the first stages in the evolution of a berry are seen in 
H. hircinum, where the fruit, although capsular and persistent, is incompletely dehiscent. In 
H. androsaemum the fruits ripen to a shiny berry (black or dark red-brown), which is easily 
detachable from the receptacle. If not eaten, after about a month they dry and can be split 
into three valves by relatively slight pressure. The seeds are then, according to Ridley (1930), 
dispersed by wind and rain. The fruits have been described as deciduous, but I have never 
seen one falling from the plant and believe that naked receptacles are always due to removal 
of the fruit by birds (cf. Robson, 1973). Ridley (1930) reported that a clump that he kept 
under observation at Kew showed no signs of having been visited by these animals, but the 
spread of H. androsaemum in New Zealand (where it was introduced) has been attributed to 
their activities (Thomson, 1922; Connor, 1977). My own observations of its propensities to 
become a garden weed in various parts of Britain suggest that it is being spread by birds, 
although I, too, have failed to observe a bird eating the fruit of this species or its hybrid with 
H. hircinum (H. x inodorum Miller). 


6. Development 


Embryology 
(a) Normal form 


The first paper completely devoted to Hypericum embryology was that by Schnarf (1914), 
whose study on H. calycinum (sect. 3. Ascyreia), H. perforatum and H. maculatum 
(sect. 9. Hypericum) was followed by those of Palm (1922) on H. japonicum (sect. 30. 
Spachium), Soueges (1925) on H. perforatum, Hoar & Haertl (1932) on some North 
American species in sects 9. Hypericum, 20. Myriandra and 30. Spachium (as well as 
Triadenum virginicum (L.) Rafin.), Souéges (1936) on H. tetrapterum (sect. 9. Hypericum), 
Swamy (1946) on H. mysurense (sect. 3. Ascyreia), Govindappa (1956) on H. japonicum, 
and Rao (1957) on ‘H. patulum’ and H. mysurense. Various sections of the genus have there- 
fore been examined; and the account given by Davis (1966) summarises the findings of these 
and other authors. There is apparently no variation that could be of taxonomic value within 
Hypericum itself, but such differences occur between the Hypericeae and other tribes. 

In Hypericum the anther is bisporangiate, whereas it is tetrasporangiate in Ploiarium 
(Prakash & Lau, 1976). The microspore tetrads are tetrahedral, isobilateral or decussate, 
and the pollen grains are 2-celled when shed. The ovule in Hypericum, as in Ploiarium 
(Prakash & Lau, 1976), is anatropous, bitegmic and tenuinucellar, with the micropyle 
formed by both integuments. The archesporial cell functions directly as the megaspore 
mother cell, and a linear tetrad develops into a Polygonum-type embryo sac.* The three 


*In contrast, the tetrad may be T-shaped in Triadenum (Myers, 1963), in genera of other Hypericoid tribes and in 
Ploiarium; and in Camellia (Theaceae), the embryo sac is of the Allium (bisporic) type (Bawa, 1970). 


126 N. K. B. ROBSON 


antipodal cells are usually ephemeral but may multiply to about seven (e.g. in H. genti- 
anoides and H. punctatum) and persist into early embryogeny. The endosperm is nuclear, 
although early cell-formation misled some authors into thinking that it was helobial (cf. 
Stenar, 1938). 

The embryogeny in the Hypericoideae, as in Ploiarum (Prakash & Lau, 1976), is Solanad 
(i.e. the terminal cell divides by a transverse wall during the second cell generation and the 
basal cell forms a several-celled suspensor), whereas in the other subfamilies it is Onagrad 
(Davis, 1966). Bugnicourt (1971) gives a detailed account of the embryogeny of H. tetrap- 
terum (sect. 9. Hypericum) and evaluates previous work on the genus. 


(b) Abnormalities 


Polyembryony occurs in Hypericum tetrapterum, where Bugnicourt (1970, 1971a) observed 
additional embryos formed from a synergid by apogamy and twin embryos produced as buds 
from the suspensor. Earlier, Noack (1939) reported that in H. perforatum aposporous 
(unreduced) embryo sacs occurred in 97% of the ovules observed, and that these unreduced 
egg-cells are occasionally fertilised. On the other hand, pseudo-polyembryony resulting from 
the concrescence of two or more ovules has also been reported in H. perforatum, as well as in 
H. maculatum and H. tetrapterum (sect.9. Hypericum) and various species in other, 
distantly related sections (Bugnicourt, 19715). 


Germination 
(a) The process 


As was shown above (p. 112), the seeds and embryos of the Hypericoideae and 
Bonnetioideae (at least those of Ploiarium) are basically similar, but they differ from those of 
the other subfamilies of Guttiferae; and similar major differences occur in the germination 
process. In those taxa with free, developed cotyledons (Hypericoideae, Bonnetioideae, 
Clusieae), the germination is epigeal, whereas in those taxa where the hypocotyl is swollen 
and the cotyledons vestigial (Moronobeoideae, Garcinieae) or the cotyledons are enlarged 
and united (Calophylloideae), the germination is hypogeal (Brandza, 1908). Clusia is some- 
what intermediate in that the hypocotyl is swollen, but the cotyledons still function. 

Brandza (1908), who made a definitive study of seed-structure and germination in the 
Guttiferae, found that in Hypericum (of which he studied numerous species) the emergence 
of the radicle is followed immediately by the development of a ring of strong root-hairs at the 
base of the quickly elongating hypocotyl. These hairs appear to serve as an anchor for the 
developing seedling. They persist for a long time, until after the first foliage leaves have 
developed. 

Brandza’s observations agree with those made by other authors (e.g. Ohlendorf (1907) on 
H. elodes), and they have been confirmed many times during this study; but his remark that 
the radicle always gives rise to the main root, although true, requires some comment. 
Brandza himself noted that, in H. perforatum and H. elodes, strong adventitious roots soon 
develop from the hypocotyl and young stem base; and, indeed, in some sections of 
Hypericum the main root withers or becomes indistinguishable from the adventitious ones. 
In sects 29. Brathys and 30. Spachium (and probably also in the whole of the Hirtella group, 
sects 17-19), however, the radicle becomes a taproot, and adventitious roots develop only in 
species with decumbent or prostrate stems (e.g. H. scioanum). 


(b) Factors influencing germination 


Experimental studies of germination in Hypericum by Kinzel (1913, 1920) showed that most 
of the seeds would germinate only in light (at 20°C). The species studied were Central 
European herbs belonging to various sections, and all showed this light requirement except 
H. humifusum in part. Seeds of this species from plants growing in loose arable soils 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 127 


germinated in the dark at 20°C, but seeds from plants of clay soils behaved as obligate light- 
germinators. Subsequent authors have confirmed the light requirement for H. perforatum 
(Gensini, 1967) and H. japonicum (sect. 30. Spachium) (Isikawa, 1962). 


Vegetative development 


As far as I am aware, there have been no studies on the vegetative ontogeny of Hypericum 
other than the paper by Zimmermann (1928) on the vegetative apex in H. hookeranum and 
H. uralum (both sect. 3. Ascyreia). On the other hand, several investigators have studied the 
ontogeny of the flower. 


Floral development 


Although the basic structure of the flower of Hypericum is best demonstrated by means of a 
study of its floral vasculature (see p. 128), the earliest floral studies were ontogenetical; and 
such studies have continued to be published over the years. 

Payer (1857) described the floral whorls in various species as appearing in acropetal 
succession, with the sepals developing quincuncially (i.e. in 2/5 spiral succession), the petals 
simultaneously, the androecial primordia shortly afterwards and, very soon after them, the 
5-2 members of the gynoecium. In species with a trimerous androecium the double stamen 
fascicles were seen to have larger primordia than the single one, and in each case the stamens 
developed in centrifugal succession. In H. prolificum (sect. 20. Myriandra) the original five 
primordia soon united in a ring, thus producing the apparently polyandrous state of the 
mature flower. 

Later authors have elaborated or disagreed with Payer’s account, but it remains essentially 
valid. Hofmeister (1868) claimed that the petal primordia appeared after those of the 
androecium; but Molly (1875), Sachs (1875), Hirmer (1917) and Leins (1964) all agreed with 
Payer. Sattler (1972) remarks that the interval between the initiation of the whorls, if it exists 
at all, is so short in H. perforatum that it could not be demonstrated with certainty. The 
difficulty, according to Molly, is that the petal primordia are soon concealed by the fascicle 
_ primordia, which develop more quickly. 

_ Breindl (1934), as we have already seen (p. 86), showed that the time and position of 
_ origin of the sepals, although basically following a 2/5 spiral succession, were modified by 
the change from foliar to calycine phyllotaxis. 

In H. elodes and H. aegypticum, Hirmer (1917) described the fasciclodes as arising long 
after the androecium had been initiated, and Leins (1964) indirectly corroborated his observ- 
ations. All accounts agree that the stamens arise centrifugally on the fascicle primordium 
(where one is visible) or directly on the torus, as they do in all closely related families; but 
' Leins (1971) and Sattler (1976) have pointed out that the direction of stamen initiation is not 
, such a fundamental character as it was thought to be by Corner (1946) and Cronquist (1968), 
for example. Indeed, Cronquist’s placing of the Myrtales in the Rosidae appears to be based 
largely on its having centripetal stamens, whereas most other considerations appear to indi- 
cate a relationship with the Guttiferae. 
| Hirmer, being unable to see five original antipetalous fascicle primordia in H. prolificum, 
| Suggested that the first groups of three stamens were antisepalous; but Payer’s observations 
are more likely to be correct in view of the structure of the flower of this species as revealed 
‘by its vasculature. In any case, Hirmer himself described the first stamen groups of H. 
_ drummondii, another ‘afascicular’ species, as antipetalous. 

| In general, then, the ontogenetic data support those from morphology in indicating a close 

relationship between the petal and the stamen fascicle interior to it. The relatively late origin 
of the fasciclodes in Hypericum is consistent with the anatomical data (see p. 140) in 
suggesting that these organs provide an example of ‘evolutionary recall’ (Robson, 1972a). 


128 N. K. B. ROBSON 
7. Floral vasculature 
Introduction 


At an early stage in these studies, it was realised that the vascular structure of the flower 
could give a clue to the basic floral structure in the Hypericoideae and its variations. 
Following earlier investigations (van Tieghem, 1875; Henslow, 1890), Saunders (1936, 1939) 
had laid the foundations of this study in Hypericum; but she confined her attentions to 
species with a fasciculate androecium. Moreover, her observation that the ‘hypogynous 
bodies’ of H. elodes and H. aegypticum were not vascularised (Saunders, 1936) conflicted 
with that of Hirmer (1917), who saw numerous ‘rudimentary’ (i.e. vestigial) vascular strands 
in them. The only reference to other species of Hypericoideae had been made by Wilson 
(1937), who described part of the floral vasculature of Vismia dealbata. The results of this 
work were presented as part of a thesis (Robson, 1956), and a summary of some aspects of it 
has been published (Robson, 1972a). 


Materials and methods 


The material was mostly reconstituted from herbarium specimens (Tillson & Bamford, 
1938), but some was fresh and was depigmented in | : 1 alcohol-acetone. For bleaching, the 
method described by Vautier (1949) was adopted at first, later being replaced by the use of 
commercial bleach, which produced quicker results. The buds were cleared in lactic acid 
(Sporne, 1948). 

The species studied were as follows (for details, see Robson [1956]): 


HYPERICUM 


Sect. 1. Campylosporus : H. revolutum. 
3. Ascyreia : H. oblongifolium, H. calycinum, H. forrestii. 
5. Androsaemum : H. x inodorum. 
7. Roscyna : H. ascyron. 
9. Hypericum : H. erectum, H. scouleri, H. punctatum. 
9a. Concinna : H. concinnum. 
10. Olympia : H. olympicum. 
14. Oligostema : H. humifusum. 
18. Taeniocarpium : H. pulchrum, H. thymifolium. 
20. Myriandra : H. prolificum, H. kalmianum, H. ellipticum, H. crux-andreae 
(H. stans (Michx.) Adams & Robson), H. hypericoides. 
21. Webbia : H. canariense. 
25. Adenotrias : H. aegypticum. 
26. Humifusoideum : H. saruwagedicum, H. natalense, H. peplidifolium. 
27. Adenosepalum : H. aethiopicum subsp. sonderi. 
28. Elodes : H. elodes. 
29. Brathys : H. goyanesii. 
30. Spachium : H. brasiliense, H. canadense, H. gentianoides. 


VISMIA 


V. magnoliifolia Cham. & Schlecht., V. crassa (Rusby) Blake, V. guianensis (Aublet) Choisy, 
V. micrantha Mart. 


CRATOX YLUM 
C. cochinchinense (Lour.) Blume. 
TRIADENUM 
T. walteri (J. F. Gmelin) Gleason. 
PLOIARIUM 


P. alternifolium (Vahl) Melch. 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 129 


Hypericum—the torus 


(a) Basic plan 


The basic plan of the toral vasculature in Hypericum is shown as a coplanar diagram in Fig. 
32. The sepal traces (S 1-5), which depart from the stele at different levels, are unilacunar 
but divide into three almost immediately. Above each of them, the lacuna is large, due to the 
absence of the antisepalous androecial whorl. The petals, likewise, have unilacunar, three- 
branched traces (P); but they all depart at the same stelar level. Above each one is a relatively 
massive stamen-fascicle trace (St), also three-branched; and the remaining stelar tissue 
vascularises the gynoecium, first branching into three to form the dorsal carpel trace (CD) 


= Se] Sen 


CV 


cD 


CL 


St 


1 ns 2 
Fig. 32 Basic (primitive?) plan of toral vasculature in Hypericum. Legend used in Figs 32-52: 
CD =dorsal carpel trace; CL=lateral carpel trace; CSL=commissural lateral trace; CV = 
ventral carpel trace; OC = outline of carpel; OTT = ovule traces; P = petal trace; S,_, = sepal trace 
(1-5); SL=lateral sepal trace; SM=sepal midrib trace; ST=stamen fascicle trace; Std= 
fasciclode (i.e. sterile fascicle) trace; VStd = vestigial fasciclode trace. 


CV 
CD 


St 


5 Es BS - Me ee . ait a Ne, | 
Fig. 33 Plan of toral vasculature of H. revolutum (sect. 1) (see Fig. 36a). 


130 N. K. B. ROBSON 


\ f 7 | me . cD : 
(2 Y; P 
Wig vd q 


ey 
Sa es emeerenn ee 
SE € 
= 
ietoienneinn 


Fig.34 Toral vasculature of H. forrestii (sect. 3): (a) from above (x 25); (b) plan (see Fig. 36b). 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 131 


and two laterals (CL). The residual tissue comprises five ventral carpel traces (CV), which 
innervate the placentae and then dichotomise (see Fig. 48). 


(b) Isomerous pentamerous flowers 


No flower was found to have all the features of this idealised vascular pattern. Hypericum 
revolutum (sect. 1. Campylosporus) very nearly has it (Figs 33, 47a), but the sepal traces are 
2-3(4)-lacunar with the laterals partly (S 1-3) or wholly (S 4, 5) commissural. On the other 
hand, in the flower of H. forrestii (sect. 3. Ascyreia) depicted in Figs 34 and 47b the sepal 
traces are all unilacunar; but the upper parts show the results of the developmental strains 
that were discussed above (p. 86), the greatest distortion being evident near sepal 5. Note 
also that there is only one true petal-trace lacuna, as four of the stamen-fascicle traces depart 
as two halves, i.e. without joining to form an arc. Not all regularly pentamerous flowers show 
such stress effects, however. In H. ascyron (Fig. 35), the sepal traces are 3-lacunar with 
commissural laterals, whereas the stamen-fascicle traces form arcs before departing. 

Another trend, namely one of relative vertical condensation, can be seen by comparing the 
lateral half-stele views of these three species (Fig. 36). This is more pronounced in the 
flowers of other sections. The other regularly pentamerous flowers (of H. oblongifolium and 
H. calycinum) had a vascular structure essentially similar to those already described. 


(c) Flowers with pentamerous androecium and trimerous gynoecium 


Where the gynoecium has become trimerous, as in Hypericum x inodorum (Fig. 37), the 
elimination of two carpels is not completely reflected in the vascular structure. The dorsal 
traces of carpel | are only slightly distorted, but those of the other two are both compound, 
including parts of the traces to the ‘missing’ carpels 4 and 5. It is clear, however, that the 
contractions have occurred along the radii of sepals 4 and 5. 


(d) Flowers with trimerous androecium and gynoecium 


In a large-flowered species with trimerous inner whorls (e.g. Hypericum olympicum, Figs 38, 
47d), all signs of carpels 4 and 5 are absent; but all five androecial traces are still present, one 
single (above petal 1/3) and the others in pairs (above sepals 4 and 5). The morphological 
evidence of union of fascicles, discussed above (p. 100), is thus reflected in the traces to these 


CV 


CL 
cD 


St 


SL 


SM 


= fe» 
Fig. 35 Plan of toral vasculature of H. ascyron (sect. 7) (see Fig. 36c). Note that stamen-fascicle 
traces are single (in fact, curved) for some distance. 


132 N. K. B. ROBSON 


cv cp ck cv CL Go 


Fig. 36 Half toral vasculature of: (a) H. revolutum; (b) H. forrestii; (c) H. ascyron (all x 20). 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 133 


Fig. 37 Toral vasculature of H. x inodorum (sect. 5): (a) from above (x 15); (b) plan. 


134 N. K. B. ROBSON 


Fig. 38 Toral vasculature of H. ol/ympicum (sect. 10): (a) from above (x 20): (b) plan. Note the 
merging of stamen-fascicle traces 2/4 and 3/5. 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 


Fig. 39 Toral vasculature of H. pulchrum (sect. 18): (a) from above (x 36); (b) plan. 


135 


136 N. K. B. ROBSON 


CV 


6 CD 


(i) (ii) (iii) iw () Wi) 


Fig. 40 Plan of toral vasculature of a 4-carpellary flower of H. natalense (sect. 26), showing that 
the trace to carpel 4 is derived from that of carpel | of a 3-carpellary gynoecium. 


organs. In smaller-flowered species (H. pulchrum (Figs 39, 47e) and H. aethiopicum subsp. 
sonderi), the vascular plan is essentially similar but simpler, as it is in the species with 
fasciclodes, H. aegypticum and H. elodes (Robson, 1972a: figs 8, 9). 


(e) Sect. Humifusoideum 


Saunders (1937) investigated Hypericum peplidifolium because it was alleged to have three 
fascicles and five styles (Keller, 1925), which seemed to go against the trend of meiomery 
starting in the innermost organ. She claimed that all the flowers that she examined had five 
fascicles, not three; but Milne-Redhead (1953) described this species as having stamens : 
‘usually in 3 often rather indefinite groups of 7-10, but sometimes totalling as many as 10’. 

H. peplidifolium is a relatively advanced, herbaceous member of sect. 26. Humifusoideum. 
When a more primitive, shrubby member (H. saruwagedicum) was studied, its toral vascula- 

ture was found to be essentially similar to that of H. goyanesii (see below, Fig. 45), i.e. with 

five distinct fascicle traces ‘spread out’ so that their branches form a more-or-less continuous | 
row. This pattern reflects that of the stamens, which are ‘irregular’ or ‘not in fascicles’. The | 
gynoecium is trimerous. 

H. natalense is intermediate between H. saruwagedicum and H. peplidifolium in having, 
for example, a 4—5-carpellary gynoecium and stamens ‘irregularly arranged in 3 or 4 
indistinct groups’ (Killick & Robson, 1976). Figure 40 shows that the grouping of the 
androecial traces is also indistinct, and that the fourth carpel is innervated by the division of 
the trace opposite sepal | (shaded area) of a 3-carpellary gynoecium. 

In H. peplidifolium (Fig. 41) this doubling process has been repeated opposite sepal 2 (Fig. 
41a) or 3 (Fig. 41b); and the androecial traces reflect the restored pentamery to a greater (Fig. 
41a) or lesser (Fig. 41b) degree. The vascular pattern of the torus therefore supports the 
hypothesis, based on gross morphological trends, that the androecium and gynoecium in 
sect. 26. Humifusoideum is secondarily pentamerous. 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 137 


Gi) wid (vili 


! 
we t TM: 
| 


Gi (iv) (v) 


a 


es ‘ 


\ 

I 

! 

| 
WER 
VE AVE. 
OE 
| 

I 

{ 


bes ; : 


diy Gi i) (vii) (vii) (ix) 


Fig. 41 Plans of toral vasculature of 5-carpellary flowers of H. peplidifolium (sect. 26), showing: 
(a) irregular stamen-fascicle traces, with traces to carpel 5 derived from those to carpel 2; more 
regular stamen-fascicle traces, with traces to carpel 5 derived from those to carpel 3. (1)-(x) = half 
carpel-traces (numbered from |. to r., i.e. numbers in (a) and (b) do not correspond). 


138 N. K. B. ROBSON 


CV 
cD 


\ st 


os2=e es, 
- =. 
= = 
~ 


Fig. 42. Toral vasculature of H. prolificum (sect. 20): (a) half (lateral view) (x 28); (b) plan. Note 
the union of stamen-fascicle traces on either side of sepal 5. 


(f) Polyandrous species 


Keller (1925) described three sections of Hypericum as having polyandrous afasciculate 
androecia, viz. Campylopus, Myriandra and Brathys. Of these, sect. 11. Campylopus (i.e. H. 
cerastoides) was found to have five fascicles of stamens which varied from free to united 
1 + 2 + 2; and, as the fascicles were all only slightly united at the base, they were easily 
separable; sect. 20. Myriandra now includes the ‘genus’ Ascyrum L.; and Brathys has proved 
to comprise two independent sections, 29. Brathys and 30. Spachium (Robson, 1977a). 

In sect. Myriandra the stamens are inserted on a relatively wide toral zone and are shed 
individually, and the androecial traces are likewise branched to cover a wide area (Figs 42, 
47g). These traces, however, are basically similar to those of fascicled species (Figs 33-39), 
except that there is greater lateral condensation and less vertical condensation. Continued 
lateral condensation is correlated with meiomery of the perianth in the species formerly 


»| 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 139 


Fig. 44 Plan oftoral vasculature of H. brasiliense (sect. 30). 


_ placed in Ascyrum; but, here too, the fascicle traces are visible. They have, of course, been 
_ reduced in number to four (Fig 43, 47h). 
In sects 29. Brathys and 30. Spachium, on the other hand, the androecial zone is relatively 
| narrow, the stamens often forming only a single row. Even so, the basic toral vasculature is 
_ typical (Figs 44, 45), with five fascicle traces clearly visible. In those species of sect. 
Spachium where the flowers are smaller and the stamens fewer, e.g. H. canadense (Figs 46, 
47k), the discontinuous ring of stamens is reflected in a less regular vascular pattern; but the 
five fascicle traces are still present. The whole flower, however, has undergone relative 
| vertical condensation in these sections, so that the stamens can be said to have been 
“squeezed out’ into a narrow ring. 


140 N. K. B. ROBSON 


bees 4 — 


Fig. 45 Plan of toral vasculature of H. goyanesii (sect. 29). Note how five carpel traces serve three 
carpels (1,2 +4, 3 +5). 


Fig. 46 Plan of toral vasculature of H. canadense (sect. 30). Note the extreme ‘dissection’ of the 
carpel traces. 


(g) The ‘Elodes’ group 


Hypericum aegypticum and H. elodes have toral vasculature that is typical of ‘3+3’ 
Hypericum species. The fasciclodes have no vascular connections with the stele (cf. Robson, 
1972 : figs 8, 9). In H. aegypticum but not H. elodes, however, there are obscure vestigial 
vascular strands within the fasciclode. 


Hypericum—the ovary 


Species of Hypericum with large ovaries have a dorsal carpel trace that branches in the ovary 
wall and near the stigma, and the lateral carpel traces penetrate to the style and dichotomise 
(Fig. 48a). In more advanced species with a relatively massive style, the branching of the 
dorsal trace may begin at a lower level (Fig. 48b); but the general trends are towards reduc- 
tion of the branching and elimination of the lateral traces (Fig. 48c, d). 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 141 
Other genera 


In contrast to the fasciclodes of the ‘Elodes’ group, those of Ploiarium, Vismia, Cratoxylum 
and Triadenum are fully vascularised (cf. Robson, 1972a : figs 5-7; Baas, 1970), their traces 
bearing the same relationship to those of the sepals as the stamen fascicle traces bear to those 
of the petals. This relationship is clear in the regularly pentamerous Vismia (Fig. 49) (cf. also 
Wilson (1937 : fig. 29)); and even in Cratoxylum, where the androecium and gynoecium are 
trimerous, there are vestigial traces to the two ‘missing’ fasciclodes (Fig. 50). The traces to the 
stamen fascicles opposite the sepals, however, are double in this genus and in Triadenum 
(Fig. 51), as they are in Hypericum. The ovary vasculature of Vismia and Cratoxylum is 
more complex than in Hypericum, (i) being more elaborately branched and (ii) having the 
lateral carpel traces always departing from the dorsal carpel trace (Fig. 52a, b). Triadenum 
shows a simplification of the Cratoxylum pattern (Fig. 52c). 


Discussion 


From the vascular patterns described above, it will be seen that the floral anatomy of 
Hypericum can be interpreted in terms of variations on a basic plan (Fig. 32). These 
variations in general parallel the morphological changes already discussed (e.g. the 
formation of ‘double’ stamen fascicles) and frequently indicate a previous evolutionary stage 
when gross-morphological evidence for this has disappeared (e.g. the 5- or 4-fascicled nature 
of the androecium in sect. 20. Myriandra). 

The patterns in Vismia and Ploiarium (not illustrated, but essentially similar to that of 
Vismia) support the idea that these genera have isomerous pentamerous diplostemonous 
androecia (i.e. that the fascicle functions as the androecial unit), and that the androecium 
becomes haplostemonous in Hypericum by suppression of the antisepalous whorl. The floral 
vascular pattern also allows discrimination between meiomery resulting from the loss of an 
organ (e.g. the perianth members of tetramerous Hypericum species or the gynoecium of 
Cratoxylum, Triadenum and some species of Hypericum) and from fusion of adjacent organs 
(e.g. the androecia of Cratoxylum, Triadenum and some species of Hypericum). Finally, the 
relationship of the fascicle or fasciclode trace to that of the sepal or petal below it is 
comparable with the trace of an axillary branch to that of its subtending leaf, thus lending 
support to the Gonophyll Theory (Melville, 1962, 1963). 

Likewise, the alternation of dorsal or dorsal-plus-lateral carpel traces (supplying the 
‘carpei’ wall and style) with ventral carpel traces to the placenta and ovules invites compari- 
son with the alternation of sepal and petal traces. In other words, the traces to the ‘carpel’ 
walls and the placentas remain independent. There is therefore no vascular evidence in 
Hypericum and related genera that favours the Carpel Theory—pace Dupuy & Guédeés 
(1975). Indeed, it seems to be special pleading to interpret the strong vasculature of the 
Hypericum placenta as related to its nutritive function alone. ‘Occam’s Razor’ would require 
one to describe the ovary of Hypericum species with a pentamerous gynoecium as com- 
prising five foliar organs on the sepal radii alternating with five, usually 2-lobed placentae on 
the petal/stamen-fascicle radii (see p. 106 and Figs 23, 24). In terms of the Gonophyll 
Theory these are five tegophylls alternating with five fertile gynophyll branches. The two 
components usually become separated in fruit, the tegophylls being easily removable, 
leaving the five or fewer 2-lobed placental branches as a 1-5-partite ‘column’. 

The evidence from floral vasculature has a bearing on two more morpho-evolutionary 
controversies: 

(i) The behaviour of the stamen fascicle and fasciclode as androecial entities (i.e. as 
equivalent to a single stamen and staminode in oligostaminal flowers) indicates that the 
individual stamens in each fascicle are to be interpreted as branches of this entity. Increase or 
decrease in stamen number and direction of maturation are thus equivalent to a variation in 
the amount of branching, not to an increase or decrease in numbers of whole organs. They 
are not, therefore, of fundamental evolutionary significance; and it would seem to be wrong 


142 N. K. B. ROBSON 


i] “ { 
| 
v Vv 


Fig. 47 Longitudinal sections of toral vasculature of Hypericum species, showing some trends in 
vertical and lateral condensation (numbers indicate sections): (a) H. revolutum (1); (b) H. forrestii 
(3); (c) H. calycinum (3); (d) H. olympicum (10); (e) H. pulchrum (18); (f) H. canariense (21); (g) 
H. prolificum (20); (h) H. hypericoides (20); (i) H. goyanesii (29); (j) H. brasiliense (30); (k) H. 
canadense (30) (a—d x 23, e-k x 28). 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 143 


144 N. K. B. ROBSON 


i) (Ts) Se 


“ae 


Fig. 48 Carpel vasculature in Hypericum (omitting ovule traces): (a) H. revolutum (x c. 10); (b) H. 
ascyron (x c. 12); (c) H. olympicum (x 14); (d) H. aethiopicum subsp. sonderi (x c. 12). 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 145 


Q = a . Q 9 2 D = 


Fig. 49 Plan of toral vasculature of Vismia guianensis, showing that the fasciclode traces (Std) 
bear the same relationship to the sepal traces as the stamen-fascicle traces (St) do to the petal 
traces. 


Fig. 50 Plan of toral vasculature of Cratoxylum cochinchinense, showing two vestigial fasciclode 
traces (V Std). X = ‘blind’ stamen-fascicle traces. 


| to apply the term ‘centrifugal’ to both the direction of maturation of stamens in a fascicle and 
obdiplostemony (Sattler, 1976; Tucker, 1976). The latter is a function of the relative time of 
maturation of whole organs, not branches. For example, the delayed development of the 
___ fasciclodes in Hypericum aegypticum and H. elodes results in this condition. 

| (11) Union of stamen fascicles and union of stamens within a fascicle are equivalent to the 
union, respectively, of organs and branches of organs. The two trends are distinct, though 
| frequently concurrent. It is doubtful if any member of the Guttiferae retains any of the 
primitive ‘trunk’ of the stamen fascicle very far above the toral level (see p. 99). In every 
case examined, e.g. Hypericum aegypticum and H. elodes (Robson, 1956), the Morono- 
beoideae (Wilson, 1937; Kawano, 1965), Cratoxylum (Baas, 1970) and Garcinia species 
(Stebbins, 1974), the traces to the individual stamens run separately through the ‘trunk’. 


146 N. K. B. ROBSON 


Fig.51 Plan of toral vasculature of Triadenum walteri. Traces marked ‘x’ serve stamen fascicles, 
not fasciclodes. 


Fig. 52 Carpel vasculature in: (a) Vismia guianensis (x 15); (b) Cratoxylum cochinchinense 
(x 9); (c) Triadenum walteri (x 9). 


8. Anatomy and phytochemistry—taxonomic implications 
Vegetative anatomy 


The vegetative anatomy of Hypericum has been studied by various workers, and summaries 
and discussions of their conclusions have been published by Vestal (1938) and Metcalfe & 
Chalk (1950). In addition, Schofield (1968) made a survey of nodal and petiolar anatomy in 
the Guttiferae and related families. More recently, Baretta-Kuipers (1976) has compared the 
wood anatomy of Hypericoid genera with that of the rest of the Guttiferae, particularly the 
Bonnetioideae. In the light of these studies, and because Dr A. C. Gibson (of the University 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 147 


of Arizona, Tucson) is engaged on a comparative anatomical study of the Cratoxyleae and 
Hypericeae, it is unnecessary to make a further detailed analysis of Hypericum anatomy 
here. 

In general, Hypericum is anatomically distinct from the genera of Vismieae and Cratoxy- 
leae. Vestal (1938), on anatomical grounds, supported Hochreutiner’s (1919) suggestion that 
it is more closely related to the Clusioideae, but that ‘the arborescent Hypericaceae’ are 
derived from the Calophylloideae. This is too simplistic a view; and to link these latter 
groups through Psorospermum, as Hochreutiner and Vestal suggested, is clearly erroneous. 
Nevertheless, Hochreutiner’s proposals serve to emphasise the distinctness of each of the 
Hypericoid tribes. Baretta-Kuipers (1976) states that these tribes ‘are all quite different [in 
wood anatomy] from the Bonnetiaceae’, but the data in her table 2 do not seem to bear this 
out. From a relatively cursory study of the literature, I would suggest that the Hypericoideae 
are anatomically more similar to the Bonnetioideae than they are to other subfamilies of the 
Guttiferae, but that they are in general most advanced than are the Bonnetioids. 


Secretory system 
(a) Introduction 


The nature and external distribution of the glands in Hypericum have been considered under 
Morphology. It is necessary here to discuss their internal distribution, their chemical 
composition and their taxonomic significance. 


(b) Distribution of secretory cavities and canals 


Secretory canals, containing essential oils, are present in Hypericum in the stem and root. 
They also penetrate, as we have seen, into the leaves, sepals and petals, where (i) they are 
frequently dissected into streaks and dots (isodiametric lacunae) and (ii) their contents are 
often denser and darker in colour (Coutinho, 1950; Mathis, 1963). Lastly, they occur in the 
ovary wall, where they may also become streaks or dots (i.e. vesicles). The external stem 
glands (when present) and the anther-connective glands are always punctiform or short, and 
their contents are resinous or waxy. 
In the vegetative parts of the plant, the glandular canals are distributed in the following 
way: 
Root — phloem, pericycle. 
Stem — medulla (occasionally one central canal, 
e.g. in H. calycinum, H. balearicum, 
H. linarifolium), phloem, pericycle, 
cortex (rarely, e.g. in H. calycinum). 
Leaf — mesophyll (often dissected), vascular bundles 
(phloem). 
The above data were obtained from Kexel (1896), Costa (1904) and Coutinho (1950). The 
secretory canals of the Cratoxyleae have a similar distribution, but seem to occur constantly 
in the stem medulla and possibly the cortex (Kexel, 1896; Holm, 1903; Baas, 1970). 


(c) Composition of contents of secretory cavities and canals 


According to Mathis (1963), the essential oils in the secretory tissues of Hypericum can be 
divided into two fractions, volatile and heavy, depending on their behaviour in chemical 
analysis. The volatile fraction, which contains saturated hydrocarbons and monoterpenes, 
comprises 60-80% of all the essential oils; and the heavy fraction contains mainly sesquiter- 
penes, monoterpene alcohols and linear-chain aldehydes. Mathis found that the shrubby 
species (in sects 3. Ascyreia, 5. Androsaemum, 6. Inodora and 20. Myriandra) were rich in 
(volatile) limonene and often in myrcene, and that there was sometimes a predominance of 
monoterpene alcohols in the heavy fraction (cf. also Mathis & Ourisson, 1964a, b). On the 
other hand, the dwarf shrubs and herbs (in sects 7. Roscyna, 9. Hypericum, 10. Olympia, 


148 N. K. B. ROBSON 


Table 6 Distribution of hypericin and pseudo-hypericin in Hypericum (modified from Mathis (1963)) 


Section Species tested (No.) Stem Leaf Sepal Petal Notes 
1. Campylosporus all (5) — + + + 
2. Psorophytum all (1) — — —_— — 
3. Ascyreia all (12) _ — — — 
4. Takasagoya all (1) — — — — 
5. Androsaemum all (4) — — — — 
6. Inodora all (1) — — — — (a) 
7. Roscyna all (2) — — — — 
8. Bupleuroides all (1) — — + + 
9. Hypericum most (15) — + + + 
others (9) + + + + (b) 
10. Olympia all (2) — + + + 
11. Campylopus all (1) —_ + + + 
12. Origanifolia all (2) + + + + 
13. Drosocarpium all (10) — + + + 
14. Oligostema all (6) — + + + 
15. Thasia all (1) — + + + 
16. Crossophyllum H. orientale — — — — (c) 
H. adenotrichum — + + + 
17. Hirtella most (10) — — + + (d) 
others (7) + + + 
H. retusum + + + + 
18. Taeniocarpium all (12) — + + (e) 
19. Coridium most (3) — — + — 
H. roberti — — + + (f) 
20. Myriandra all (16) — — — — 
21. Webbia all (1) — — — — 
22. Arthrophyllum all (3) — — — — (g) 
23. Triadenioides H. scopulorum — —_— + + 
H. pallens — + + + 
24. Heterophylla — 
25. Adenotrias all (2) — — — — 
26. Humifusoideum all (1) — + + + 
27. Adenosepalum most (16) = + + + 
H. aethiopicum + + + + (h) 
28. Elodes all (1) — — — — (i) 
29. Brathys all (11) — — — — 
30. Spachium all (33) — — — —_ 


Notes: (a) H. xylosteifolium has two forms, respectively without and with glandular-marginal sepals. Mathis cites a 
positive record by R. Salgues, which could be from the glandular form. (5) All positive records for hypericin in the 
stem, except that for H. punctatum Lam., come from species in the H. perforatum group. The failure of Mathis to 
record them in H. corsicum Steudel (=H. tetrapterum) may be due to a misidentification. (c) Of the two species in 
this section, H. orientale has only amber glands and H. adenotrichum has black glands. (d) The ten species include 
three (H. amanum Boiss.=H. amblysepalum, H. leptocladum Boiss.=H. helianthemoides, H. assyriacum 
Boiss. = H. hirtellum var.) in which presence in the leaves is listed as doubtful. (e) H. fragile, which is stated to have 
hypericin in the leaves, belongs to this section; but cultivated material thus named is nearly always H. olympicum 
(sect. 10. Olympia). (f) H. roberti is a synonym of H. ericoides, in which the petal margins vary from eglandular to 
black-glandular. (g) H. rupestre has glandular-margined sepals but apparently gave a negative result. (h) H. 
aethiopicum subsp. sonderi (Bredell) N. Robson has a black-glandular stem; in the other subspecies, and in the other 
species listed, it is eglandular. (i) The red marginal glands of the sepals apparently do not contain enough hypericin 
to give a positive result. 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 149 


13. Drosocarpium, 18. Taeniocarpium, 19. Coridium and 27. Adenosepalum) were poor in 
limonene and myrcene, and the heavy fraction from them consisted mainly of sesquiterpenes 
and sometimes of saturated linear-chain aldehydes (cf. also Mathis & Ourisson, 1964c, d). 

On the basis of Mathis’s data, it does not seem possible to make any taxonomically more 
useful generalisations. He found quantitative seasonal variation, as well as variation with 
age. Thus the volatile fraction is most abundant (1) at the time of flowering (except in H. 
calycinum, where the fruits are particularly rich in volatile substances) and (11) in young 
plants (and presumably young parts of plants). The viscosity of the essential oils was 2-10 
times as great in the leaves as in the flowers and fruits; and the viscosity of the leaf extract was 
yellowish, whereas that of the flowers and fruits was almost colourless. The chemical 
composition of the essential oils in the different parts, however, scarcely varied. 


(da) Distribution of hypericin and pseudo-hypericin 


This topic has already been discussed on the basis of presence or absence of dark glands (Fig. 
12, p. 82). Mathis & Ourisson (1963) and Mathis (1963) gave the results of a comprehensive 
investigation of the distribution of these substances, both within the plant and throughout 
the genus. Their results were expressed in terms of Keller’s classification and included 
references to previously published results. Where these differ from the results of Mathis, they 
would appear to be due, in most cases, to misidentification or faulty technique. I have not 
therefore taken these aberrant records into account in Table 6, except in note (a). 

Table 6, in which Mathis’s results have been listed according to my classification (Robson, 
1977a), should be compared with Fig. 12. It will be seen that the occurrence of black glands 
in an organ is an accurate indication of the presence of hypericin (and probably pseudo- 
hypericin) in that organ, but that, if the glands are red (as in the sepals of H. elodes), the 
substances may not be detectable by the methods adopted. 

It should also be noted that Mathis did not test Hypericum ovaries, some of which also 
contain black glands, e.g. that of H. richeri (sect. 13. Drosocarpium). 


(e) Chemistry of hypericin and pseudo-hypericin 


Hypericin and pseudo-hypericin, which are naphtho-dianthrones related to emodin, have 
been recorded only from the Hypericeae. It is not known whether the contents of the dark 
glands in the Vismieae and Cratoxyleae are chemically identical or not; but the bark of 
Harungana madagascariensis Lam. ex Poiret (Vismieae) was found to contain a related 
substance, harunganin, which, like hypericin and pseudo-hypericin, has a molecule with the 
anthranol type of structure (Stout et al., 1962). The molecules of hypericin and pseudo- 
hypericin are very similar, differing only in one radicle (Fig. 53). 


OH OO OH CU dOwh eH 
we OH — CHOHCH; OH 
CH, OH  CHOH-CH, on 
OH O OH OH O OH 
a b 


Fig.53 Molecular structure of: (a) hypericin; (b) pseudo-hypericin (after Mathis, 1963). 


150 N. K. B. ROBSON 
Chemotaxonomy of the Hypericoideae 


A short summary of chemotaxonomical data on the Hypericoideae, most of which refers to 
Hypericum, was given by Hegnauer (in Robson, 1974). As yet, none of it is relevant to the 
infrageneric classification of Hypericum. The presence of xanthones, however, links the 
Hypericoideae to the rest of the Guttiferae, including the Bonnetioideae, and differentiates 
the family from the Theaceae (cf. Kubitzki, Mesquita & Gottlieb, 1978; Gunatilaka, 
Balusubramian & Kumar, 1979). 

Webb (1980), who studied the flavonols and flavonoids in the leaves of the eastern North 
American species of sect. 30. Spachium, found that each of the eleven species had a 
distinctive chromatographic profile and a characteristic set of compounds. 


9. Cytology and genetics 


Chromosomes of Hypericum 
(a) Chromosome numbers 


In a previous paper (Robson & Adams, 1968), a summary was given of the information then 
available about the chromosome numbers in Hypericum and other genera of the Guttiferae. 
(As far as I am aware, no counts have been made for species of the Bonnetioideae.) It was 
concluded that the basic numbers in Hypericum form a descending series from 12 to 7, witha 
possible extension to 6 if the count of 2n=24 for H. gentianoides proved to indicate tetra- 
ploidy. No count based on n = 11 had then been made. Tetraploidy had been recorded on the 
base numbers n = 8, 9 and 10, but not on n=7 or definitely on n= 12; and higher degrees of 
polyploidy appeared to be confined in nature to sect. 9. Hypericum and were associated with 
the largely apomictic H. perforatum (2n = 32, 48) and its hybrid with the tetraploid sub- 
species of H. maculatum (H. x desetangsii Lamotte nm. desetangsii) (2n = 32, 40, 48). 

Counts published subsequently, along with some hitherto unpublished ones made by Dr 
Mary Gibby, are incorporated in Table 7, which thus includes all the chromosome numbers 
of Hypericum that are known to me at present. Where the correct name is known, it appears 
in the second column, asterisked if the identification is doubtful. The third column includes 
names cited in the original publication where they differ from the accepted ones, due to 
synonymy or misidentification. In the next two columns, asterisked numbers are regarded as 
wrong or requiring verification. Details in the remaining columns are given only for the 
counts not cited in Robson & Adams (1968). 

When the haploid numbers (n) have been inserted in the evolutionary diagram (Fig. 54), it 
is clear that the additional records all support the general conclusions in our earlier paper 
(Robson & Adams, 1968) that: (i) the basic numbers form a descending series from 12 to 7; 
(11) polyploids occur on all these basic numbers except 7; (ili) the level of polyploidy is not 
higher than 4x, except in sects 3. Ascyreia and 9. Hypericum, in both of which it reaches 6x. 
The doubt about the ploidy of H. gentianoides (n = 12) was due to its advanced morphology; 
but similar counts for related species make it very likely that 12 is 2x, not 4x. The number 
n= 6 has, however, recently been recorded in sect. 30. Spachium (see p. 166). 

The new counts also indicate the rarity of the base number n= 11, which so far has been 
found only in sect. 3. Ascyreia. In sects 1. Campylosporus, 26. Humifusoideum and 30. 
Spachium, however, some species have n = 12 and others n = 10-7; and so the basic number 
n=11 may well be present in some or all of them, and possibly in sect. 29. Brathys. 

Polyploidy seems to be frequent in sect. 3. Ascyreia, where H. oblongifolium has both 
diploid (n= 12) and tetraploid (n=24, 22) forms; and the only count of H. monogynum 
(n=21), from a cultivated source, suggests crossing between plants with n=20 and n=22. 
The groups of Chinese and Himalayan species related respectively to H. kouytchense and H. 
patulum both seem to be basically tetraploid (n = 18); but H. augustinii, which belongs to the 
H. patulum group, is approximately hexaploid (2n = 54), the actual counts having varied 
from 2n = 54-59. Although Dr Gibby’s count of 2n = 50 for H. x moseranum (H. patulum 
(n = 18) x H. calycinum (n= 10)) is not clearly interpretable, it makes Sugiura’s earlier count 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 151 


Fig.54 Distribution of chromosome numbers in Hypericum. 


NOTE: in sect. 3 there is a secondary basic number (21); in sect. 17, x=8, 7 may occur, or there may be an 
ascending series on x’=12-®14; in sect. 28, n=10 has been recorded. For sections marked*, see 
text. 


of 2n = 36 for this artificial hybrid even more likely to have resulted from misidentification. 
It is possible that his count for H. patulum itself is also based on a misidentification. 

Likewise, Dr Gibby’s count of 2n = 30 for H. beanii ‘Gold Cup’, which has been grown at 
Hidcote Manor, Gloucs. under the name H. ‘Lawrence Johnston’, is at variance with 
Thomas’s (1970) count of n= 18 for H. patulum ‘Gold Cup’ (i.e. for the same taxon). The 
latter count may therefore have been based on a misidentification. This is especially likely as 
H. beanii ‘Gold Cup’ is the best candidate, on morphological grounds, for the role of one 
parent of H. x ‘Hidcote’, H. calycinum being the other (see p. 170). Dr Gibby’s count would 
support this hypothesis, i.e. H. beanii ‘Gold Cup’ (2n = 30) x H. calycinum (2n = 20)—-H. x 
*“Hidcote’ (2n = 50) (actual counts were c. 48, c. 54). 

In sect. 9. Hypericum both autopolyploidy (H. maculatum subsp. obtusiusculum, n= 16) 
and suspected allopolyploidy (H. perforatum, n=16) are found (Robson, 1958a), and a 
further count of 2n = 32 for H. undulatum provides additional evidence that this species does 
occur in diploid and tetraploid forms. Culwell (1970) observed lagging chromosomes, and 
consequent variation in haploid number, in H. punctatum and H. pseudomaculatum, a 
phenomenon that had already been reported in H. perforatum (Nielsen, 1924) and would 
explain Bell’s (1965) record of n=7 for H. punctatum. 

The counts of species in sect. 17. Hirtella recently made by Reynaud (1973, 1975, 1980) 
have revealed what appears to be an ascending series of diploid numbers from 2n = 20 (H. 


152 


N. K. B. ROBSON 


Table 7 Chromosome numbers in Hypericum 


Section 


Taxon 


1. Campylosporus 4H. revolutum Vahl 


2. Psorophytum 


3. Ascyreia 


4. Takasagoya 


5. Androsaemum 


subsp. keniense (Schweinf.) 
N. Robson 

H. revolutum Vahl 
subsp. revolutum 

H. balearicum L. 


H. oblongifolium Choisy 
H. oblongifolium Choisy 
H. oblongifolium Choisy 
H. monogynum L. 
H. calycinum L. 
H. dyeri Rehder 
H. uralum Buch. -Ham. ex D. Don 
H. uralum Buch. -Ham. ex D. Don 
H. x ‘Rowallane’ 
(H. hookeranum x leschenaultii) 
H. kouytchense H. Lév. 
H. kouytchense H. Lév. 
H. beanii N. Robson 
H. beanii ‘Gold Cup’ 
H. beanii ‘Gold Cup’ 
H. patulum Thunb. ex Murray 
H. forrestii (Chittenden) N. Robson 
H. forrestii (Chittenden) N. Robson 
H. forrestii (Chittenden) N. Robson 
H. x moseranum André 
(H. patulum x calycinum) 
H. x moseranum André 
H. x ‘Hidcote’ 
(parentage unknown) 
H. x ‘Hidcote’ 
H. augustinii N. Robson 


H. grandifolium Choisy 

H. grandifolium Choisy 

H. grandifolium Choisy 

H. hircinum L. 

H. hircinum L. 

H. hircinum L. 

H. hircinum var. pumilum hort. 


Original 
determination 


H. keniense 


H. cernuum 
H. cernuum 
H. cernuum 
H. chinense 


H. lysimachioides 
H. patulum 
H. patulum 


H. patulum ‘Sungold’ 
H. patulum var. henryi 
H. patulum ‘Gold Cup’ 
H. ‘Lawrence Johnston’ 


H. patulum var? 
H. patulum 


H. inodorum 


24 
22 
21 
10 
10 
10 
10 


18 
18 
ches 
18* 
18 
18 


18* 


18 


20 


2n 


24 


24 
24 


24 


20 


20 
c.36 


30 
36" 


38 


50 


c.48,c.54,55 
c.54 


40 
40 
40 


40 
40 
40 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 


Authority 


I. & O. Hedberg (1977) 


Gibby 
Nilsson & Lassen (1971) 


R.(obson)& A.(dams) (1968) 
Sareen, Kant & Pratap (1974) 
Mehra & Sareen (1969) 
Sareen, Kant & Pratap (1974) 
R. & A. 

Mehra & Sareen (1969) 
Mehra & Sareen (1969) 
Sareen, Kant & Pratap (1974) 


Gibby 

Gibby 

Thomas (1970) 
Thomas (1970) 
Thomas (1970) 
Gibby 

R. & A. 
Thomas (1970) 
Thomas (1970) 
Gibby 


R. & A. 
Gibby 


Thomas (1970) 
Gibby 
Gibby 


R.& A. 

Borgen (1969) 

van Loon (1974) 

R. & A. 

Reynaud (1975) 

van Loon & de Jong (1978) 
R.& A. 


Source 


Kenya (Mt. Elgon) 


Malawi (Mt. Zomba) 
Mallorca 


India (Chandigarh) 
India (Naintal) 
India (Chandigarh) 


India (Mussoorie) 
India (W. Himalaya) 
India (W. Himalaya) 


Chelsea Physic Garden 
Chelsea Physic Garden 

Univ. of Alabama Arboretum 
Arnold Arboretum 

Univ. of Alabama Arboretum 


Hidcote Manor (Gloucestershire) 


Arnold Arboretum 
Arnold Arboretum 
Chelsea Physic Garden 


Chelsea Physic Garden 
Armmold Arboretum 


Chelsea Physic Garden 
Chelsea Physic Garden 


ex Canary Is. (Tenerife) 
Canary Is. (Tenerife) 


Lebanon (Massif du Bedar) 
Corsica (Ghisoni) 


Location of 
voucher 


UPS 
BM 
LD 


PANJAB 
PANJAB 
PANJAB 


PANJAB 
PANJAB 
PANJAB 


BM 
BM 


BM 


MARS 


153 


154 
Table 7 (contd.) 


Section 


N. K. B. ROBSON 


Taxon 


5. Androsaemum 
(contd.) 


6. Inodora 


7. Roscyna 
8. Bupleuroides 


9. Hypericum 


H. x inodorum Miller 
H. androsaemum L. 
H. androsaemum L. 
H. androsaemum L. 


H. xylosteifolium 
(Spach) N. Robson 


H. ascyron L. 


H. nagasawai Hayata 
H. kamtschaticum Ledeb. 
var. senanense (Maxim.) 
Y. Kimura 
H. erectum Thunb. ex Murray 
H. graveolens Buckley 
H. graveolens Buckley 
H. pseudomaculatum Bush 
H. pseudomaculatum Bush 
H. mitchellianum Rydberg 
H. mitchellianum Rydberg 
H. punctatum Lam. 
H. oaxacanum R. Keller* (nomen) 
H. scouleri Hooker 
H. maculatum Crantz 
subsp. maculatum 
subsp. maculatum 
subsp. maculatum 
subsp. maculatum 
subsp. maculatum 
H. maculatum 
subsp. obtusiusculum 
(Tourlet) Hayek 
subsp. obtusiusculum 
H. x desetangsii Lamotte 
(H. maculatum x perforatum) 
nm. desetangsii 
nm. desetangsii 
H. perforatum L. 
H. perforatum L. 


Original n 

determination 

H. elatum 20 
20 


H. inodorum 
20 


18* 


8(ring) 


8(ring),7 
H. formosum 8 
H. formosum var. scouleri 8 


8 
H. quadrangulum 8 
H. fallax 
8 
H. erosum 
H. desetangsii 
16,17,18 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 


155 


ES 


2n Authority Source Location of 
voucher 
40 R. & A. 
40 R. & A. 
40 van Loon & de Jong (1978) Ireland (Roscommon) U 
40 Lessani & Chariat- 
Panahi(1979) Iran (Mazanderan) VIL 
R.& A. 
R.& A. 
Hsu (1968) 
16 R.& A. 
16 R. & A. 
R.& A. 
16 Culwell (1970) U.S.A. (N. Carolina) NCU 
R.& A. 
16 Culwell (1970) U.S.A. (Missouri) NCU 
R. & A. 
16 Culwell (1970) U.S.A. (N. Carolina) NCU 
R. & A. 
R. & A. 
Kyhos (1967) U.S.A. (California) DS 
16 R. & A. 
16 Noack (1939) 
16 Schwarz (1965) JE 
Laane (1969) Norway (Vestfold) BG 
16 Reynaud (1975) France (Pyr. Or.) MARS 
32 R. & A. 
16* Schwarz (1965) JE 
32,40,48 R.& A. 
16* Schwarz (1965) JE 
32,48 R.& A. 
32 Schwarz (1965) JE 


156 
Table 7 (contd.) 


Section 


N. K. B. ROBSON 


Taxon 


Original 
determination 


9. Hypericum 
(contd.) 


9a. Concinna 


10. Olympia 


H. perforatum L. 
H. perforatum L. 
H. perforatum L. 
H. perforatum L. 


H. perforatum L. 
H. perforatum L. 
H. perforatum L. 
H. perforatum L. 
H. perforatum 
var. microphyllum DC. 
H. perforatum 
var. angustifolium DC.* 
H. undulatum Schousb. ex Willd. 
H. undulatum Schousb. ex Willd.* 
H. undulatum Schousb. ex Willd.* 
H. undulatum Schousb. 
ex Willd. 
H. tetrapterum Fries 
H. tetrapterum Fries 
H. tetrapterum Fries 
H. tetrapterum Fries 
H. tetrapterum Fries 
H. tetrapterum Fries 
H. triquetrifolium Turra 
H. triquetrifolium Turra 
H. elegans Stephan ex Willd. 


H. concinnum Bentham 


H. olympicum L. 
H. olympicum L. 


H. olympicum L. 

H. olympicum L. 

H. polyphyllum Boiss. & Bal. 
subsp. polyphyllum 
subsp. polyphyllum 


H. veronense 


H. boeticum 


H. acutum 
H. acutum 


H. quadrangulum 


H. polyphyllum 


H. olympicum 


12 
16 


00 0O CO 


16 
16 


subsp. auriculatum 
Robson & Hub.-Mor. 


11. Campylopus H. cerastoides (Spach) N. Robson 


2n 


32 
32 
32 
32 


32 
32 
32 
32 


16* 


16 
16,32 


32 
16 


16 
16 
16 
16 
16 
16 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 


Authority Source 


Gadella & Kliphuis (1966) Netherlands (Utrecht) 
Gadella & Kliphuis (1970) Italy (Aosta) 
Nilsson & Lassen (1971) Mallorca 


Holub, Mesicek & Javurkova Czechoslovakia (Prague) 
(1972) 


Reynaud (1973) Turkey (Izmir) 
Love & Kjellqvist (1974) Spain (Jaén) 
van Loon & de Jong (1978) Yugoslavia (Zadar-Rijeka) 


van Loon & Snelders (1979) |= Greece (Athos Peninsula) 


Schwarz (1965) Austria (Graz)? 
R. & A. 

R. & A. 

R. & A. 

Sugiura (1940) 


Love & Kjellqvist (1974) Spain 
R. & A. 

Stenar (1938) 

Noack (1939) 

Polya (1950) 

Gadella & Kliphuis (1966) 
Zhukova (1967) 

Reynaud (1973) 

Reynaud (1975) 

R. & A. 


Berlin Bot. Garden 


Netherlands (Utrecht) 


Turkey (Side) 
Greece (Lidhorikion) 


R.&A. 


R. & A. 

Contandriopoulos & Lanzalavi 
(1968) 

van Loon & de Jong (1978) 

R.& A. 


Greece (Kavalla-Thessaloniki) 


Reynaud (1973) Turkey (Antalya) 


Reynaud (1973) Turkey (Antalya) 


Krusheva (1975) Bulgaria (Mt. Pirin) 


157 


Location of 
voucher 


SOA 


158 N. K. B. ROBSON 
Table7 (contd.) 


Section Taxon Original n 
determination 


12. Origanifolia H. origanifolium Willd. 9 
H. origanifolium Willd. 
H. aviculariifolium Jaub. & Spach 
subsp. aviculariifolium 
H. aviculariifolium 
subsp. depilatum (Freyn & Bornm.) 
N. Robson 


13. Drosocarpium H. vesiculosum Griseb. 
H. barbatum Jaca. H. barbatum var. typicum 


H. barbatum 
var. trichanthum (Boiss. & Sprun.) 
Boiss. 
H. barbatum 
var. pindicola Hausskn. 
H. barbatum 
var. epiroticum Hal. 
H. aucheri Jaub. & Spach 


H. rumeliacum Boiss. H. apollinis 

H. rumeliacum Boiss. if 
H. rumeliacum Boiss. H. apollinis 

H. rumeliacum Boiss. H. apollinis 

H. bithynicum Boiss. 

H. richeri Vill. subsp. richeri 7 


H. richeri Vill. subsp. richeri 
H. perfoliatum L. 

H. montbretii Spach 

H. spruneri Boiss. 


14. Oligostema H. linarifolium Vahl 
H. linarifolium Vahl 
H. humifusum L. 8 
H. humifusum L. | 


15. Thasia 
16. Crossophyllum H. orientale L. 8 

| 
17. Hirtella H. hyssopifolium Chaix H. hyssopifolium | 


subsp. hyssopifolium / 
H. lydium Boiss. / 
H. lydium Boiss. 
H. scabrum L. 
H. hirtellum (Spach) Boiss. 14 
H. helianthemoides (Spach) Boiss. 
H. vermiculare Boiss. & Hausskn. 


2n 


20 
24 
24 
24 
28 
28 
28 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 


Authority 


Source 


R. & A. 
Reynaud (1973) 


Reynaud (1973) 


Reynaud (1973) 


Reynaud (1980) 
Contandriopoulos 
& Lanzalavi (1968) 


Contandriopoulos 

& Lanzalavi (1968) 
Contandriopoulos 

& Lanzalavi (1968) 
Reynaud (1980) 
Reynaud (1973) 
Contandriopoulos 

& Lanzalavi (1968) 
R. & A. 
Reynaud (1975) 
Reynaud (1980) 
Reynaud (1973) 
R.& A. 
Reynaud (1975) 
Reynaud (1975) 
Reynaud (1973) 
Reynaud (1980) 


R. & A. 
van Loon & de Jong (1978) 
R. & A. 
van Loon & de Jong (1978) 


R. & A. 


Reynaud (1975) 
Reynaud (1975) 
Reynaud (1980) 
Reynaud (1973) 
Reynaud (1980) 
Reynaud (1980) 
Reynaud (1980) 


Turkey (Igel, Konya) 


Turkey (Antalya) 


Turkey (Icgel, Konya) 


Greece (Mt. Tymphrestos) 


Greece (Col de Peiria) 
Turkey (Kaz Dag) 


Greece (Gamila) 
Greece (Mt. Vardoussia) 


Greece (Mt. Parnassus) 
Turkey (Bursa) 


France (Hte. Loire, Alpes Mar.) 


France (Ile de Port-Cros) 
Turkey (Icel) 
Greece (Hepatochorion) 


Portugal (S. da Gardunha) 


Portugal (S. da Estrella) 


France (Var, Massif Central) 
Turkey 

Turkey (Amanus Mts) 
Turkey (Icel, Konya) 

Iran (Koram Abad road) 
Iran (Col Singal) 

Iran (Col de Assad Abad) 


Location of 
voucher 


MARS 


MARS 


MARS 
MARS 


MARS 
MARS 


MARS 


MARS 
MARS 
MARS 


MARS 
MARS 
MARS 
MARS 


MARS 
MARS 
MARS 
MARS 
MARS 
MARS 
MARS 


159 


160 
Table7 (contd.) 


Section 


N. K. B. ROBSON 


Taxon 


Original 
determination 


18. Taeniocarpium 


19. Coridium 


20. Myriandra 


H. hirsutum L. 
H. hirsutum L. 
H. hirsutum L. 
H. kotschyanum Boiss. 
H. confertum Choisy 
subsp. confertum 
H. confertum 
subsp. stenobotrys 
(Boiss.) Holmboe 
H. linarioides Bosse 


H. pulchrum L. 
H. pulchrum L. 
H. pulchrum 
forma procumbens Rostrup 
H. nummularium L. 
H. nummularium L. 
H. saxifragum 
Robson & Hub.-Mor. 


H. empetrifolium Willd. 
H. amblycalyx Coust. & Gand. 


H. coris L. 
H. coris L. 


H. kalmianum L. 
H. kalmianum L. 
H. densiflorum Pursh 
H. lobocarpum Gatt. 


H. prolificum L. 

H. frondosum Michaux 

H. galioides Lam. 

H. lissophloeus P. Adams 

H. fasciculatum Lam. 

H. brachyphyllum 
(Spach) Steudel 

H. reductum P. Adams 

H. exile P. Adams 

H. myrtifolium Lam. 

H. nudiflorum Michaux 

H. cistifolium Lam. 

H. dolabriforme Vent. 


H. densiflorum 
var. lobocarpum 


\o 


O00 0 0 O 


© 000 0 OO 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 161 


2n Authority Source Location of 
voucher 
18 R. & A. 
18 Noack (1939) 
18(+0-1B) Ferakova (1972) 
18 Reynaud (1973) Turkey (Icel) MARS 
18 Reynaud Turkey (Bursa) MARS 
18 Reynaud Turkey (Icel, Konya) MARS 
18 Contandriopoulos 
& Lanzalavi (1968) 
18 R. & A. 
18 Noack (1939) 
18 R. & A. 
18 Reynaud (1975) Spain (Pyrenees) MARS 
16 Caraz-Billat (1968) Spain (Pyrenees) 
18 Reynaud (1973) Turkey (Antalya) MARS 
18 Contandriopoulos 
& Lanzalavi (1968) 
18 Contandriopoulos 
& Lanzalavi (1968) 
R. & A. 
18 Reynaud (1975) Italy (Liguria) MARS 
R.& A. 
18 Pringle (1976) U.S.A. (Michigan) HAM 
R.& A. 
R.&A 


Mn aA ee Oe ee 
R 
> 


162 N. K. B. ROBSON 
Table 7 (contd.) 


Section Taxon Original n 
determination 


20. Myriandra H. adpressum Barton 9 
(contd.) H. adpressum 
var. spongiosum B.L. Robinson 9 
H. ellipticum Hooker 9 
H. ellipticum Hooker 
H. microsepalum 
(Torr. & Gray). A. Gray 
ex S. Watson. 
H. crux-andreae (L.) Crantz H. stans 
H. tetrapetalum Lam. 
H. hypericoides (L.) Crantz 
H. hypericoides H. stragulum 
subsp. multicaule 
(Michaux ex Willd.) N. Robson 9 
H. suffruticosum 
Adams & Robson 9 
H. x arnoldianum Rehder 
(H. lobocarpum x? densiflorum) 9 
H. x dawsonianum Rehder 
(H. lobocarpum x prolificum) 9 


\O 0 OO 


21. Webbia H. canariense L. 
H. canariense L. 


22. Arthrophyllum 

23. Triadenioides H. pallens Banks & Solander 

24. Heterophylla 4H. heterophyllum Vent. 

25. Adenotrias H. aegypticum L. 10 


H. russeggeri (Fenzl) R. Keller 10 
H. russeggeri (Fenzl) R. Keller 10 


26. Humifusoideum H. saruwagedicum Diels H. macgregorii 
H. peplidifolium A. Rich. | 


27. Adenosepalum H. glandulosum Aiton 
H. reflexum Lf. 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 163 


2n Authority Source Location of 
voucher 
R. & A. 
R.& A. 
R. & A. 
c.18 Gillett (1975) Canada (Quebec) CAN 
18 
R. & A. 
R. & A. 
R.& A. 
R. & A. 
40 R. & A. 
40 Borgen (1969) Canary Is. (Tenerife) O 
16 Reynaud (1973) Turkey (Konya) MARS 
18 Reynaud (1973) Turkey (Konya) MARS 
Ornduff in Robson (1977a) Morocco UC 
Reynaud (in litt. 1978) Syria (Herife) MARS 
Reynaud (1980) Syria (Slenfé) MARS 
24 R. & A. 
16 I. & O. Hedberg (1977) Ethiopia (Gulama Mts.) UPS 
18 van Loon & de Jong (1978) Canary Is. (Tenerife) 


U 
18 Borgen (1969) Canary Is. (Tenerife) O 


164 
Table 7 (contd.) 


Section 


N. K. B. ROBSON 


Taxon 


27. Adenosepalum 
(contd.) 


28. Elodes 


29. Brathys 


30. Spachium 


H. reflexum 

var. lanuginosum Pitard 
H. coadunatum Chr. Sm. 
H. coadunatum Chr. Sm. 
H. pubescens Boiss. 
H. tomentosum L. 
H. delphicum Boiss. & Heldr. 
H. caprifolium Boiss. 
H. annulatum Moris 
H. montanum L. 
H. montanum L. 
H. montanum L. 
H. atomarium Boiss. 
H. lanuginosum Lam. 

var. lanuginosum 

var. lanuginosum 
H. lanuginosum 

var. scabrellum 

(Boiss.) N. Robson 

H. huber-morathii N. Robson 
H. elodeoides Choisy 


H. elodes L. 
H. elodes L. 
H. elodes L. 


H. stenopetalum Turcz. 
H. campestre Cham. & Schlecht.? 


H. drummondii 
(Grev. & Hook.) Torr. & Gray 


Original 
determination 


H. irazuense 


H. sp. aff. paniculatum 


H. gentianoides (L.) Britt., Sterns & Poggenb. 
H. gentianoides (L.) Britt., Sterns & Poggenb. 


H. denticulatum Walter 
H. denticulatum Walter 
var. denticulatum 


H. denticulatum Walter 
var. acutifolium (Ell.) Blake 


H. harperi R. Keller 
H. setosum L. 
H. cumulicola (Small) P. Adams 
H. anagalloides 
Cham. & Schlecht. 
H. majus (A. Gray) Britton 
H. majus (A. Gray) Britton 
H. canadense L. 
H. canadense L. 
H. canadense L. 
H. gymnanthum Engelm. & Gray 


co 00 CO CO CO 


2n 


18 
18 
18 
18 


16 
16 
16 
16 
16 
16 
16 


16 
a2 


16 
16 


20 


24+ 


16 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 


Authority 


Source 


van Loon & de Jong (1978) 
R. & A. 

Ortega & Navarro (1978) 
K. Jones (pers. comm.) 1980 
R. & A. 

Reynaud (1980) 

Love & Kjellqvist (1974) 
Reynaud (1980) 

R. & A. 

Noack (1939) 

Reynaud (1975) 

Reynaud (1973) 


Reynaud (1973) 
Reynaud (1980) 


Reynaud (1973) 
Reynaud (1980) 
R. & A. 


R.&A. 
Dehay (1972) 
Gibby 


R.& A. 
Huyhn (1965) 


Webb (1980) 


R. & A. 
Webb 
R. & A. 


Webb (1980) 


Webb (1980) 


Webb (1980) 
Webb (1980) 
Webb (1980) 


Gillett (ined.) 
R. & A. 
Webb (1980) 
R. & A. 
Webb (1980) 
Moore (1973) 
Webb (1980) 


Canary Is. (Tenerife) 


Canary Is. (Gran Canaria) 
R.B.G., Kew 


Greece (Mt. Dirphys) 


Spain (Jaen) 
Greece (Col de Peiria) 


France (Hte Loire) 
Turkey (Izmir) 
Turkey (Antalya) 
Syria (Slenfé) 


Turkey (Antalya) 
Turkey (Korkuteli to Emali) 


France (Brenne) 
Chelsea Physic Garden 


Peru (Cord. de Vilcabamba) 


U.S.A. (Alabama, Mississippi, 
Tennessee) 


U.S.A. (Alabama, N. Carolina) 


U.S.A. (New Jersey, N. Carolina 


Tennessee) 


U.S.A. (Alabama, N. Carolina, 


Tennessee) 
U.S.A. (Florida, Georgia) 


U.S.A. (N. Carolina, S. Carolina) 


U.S.A. (Florida) 


Canada (Vancouver I.) 


165 


Location of 
voucher 


BM 


NEU 


TENN 


TENN 


TENN 


TENN 
TENN 
TENN 
TENN 


DAO 


U.S.A.(Vermont), Canada(Quebec) TENN 


U.S.A. (New Hampshire, Maine) 


Eire (Co. Mayo) 
U.S.A. (Alabama, Mississippi, 
N. Carolina) 


TENN 


TENN 


166 N. K. B. ROBSON 
Table7 (contd.) 


Section Taxon Original n 
determination 


H. mutilum L. 
H. mutilum L. 
H. boreale (Britton) Bickn. 
H. boreale (Britton) Bickn. 
H. boreale (Britton) Bickn. 
H. japonicum 

Thunb. ex Murray 8 
H. scioanum Chiov. 
H. gramineum G. Forster 8 
H. gramineum G. Forster H. sp. 


CO CO CO CO 


*Identification or count doubtful. 
R. & A. Robson & Adams (1968). 
+2n=24 in R. & A. was an error for n= 24. 


hyssopifolium), through 2n = 24 (H. lydium, H. scabrum) to 2n = 28 (H. helianthemoides, H. 
vermiculare, H. hirtellum). To read the series as a descending one would be inconsistent with 
morphological and distributional evidence, and it seems unlikely that 24 is a triploid based 
on 8 and 28 a tetraploid based on 7. Perhaps the occurrence of irregular pollen grains in this 
section (Fig. 29, p. 119; see also p. 117 and Clarke, 1975) is an indication of abnormalities in 
the breeding system in sect. Hirtella. If it is so, then we may have an explanation of some of 
the taxonomic difficulties in this section. 

The only other count that does not appear to fit into the general evolutionary hypothesis is 
that of n= 10 for H. elodes by Dehay (1972); but Dr Gibby has confirmed the occurrence of 
this number by a mitotic count. My own record (n = 16) agrees better with the interpretation 
(based on morphology and distribution) of H. elodes as a derivative of sect. 27. Adeno- 
sepalum (n=9, 8); but an increase in basic number from 9 to 10 would not seem to be 
impossible. Only further counts in sects 27 and 28 will solve this problem. 

Webb (1980) has recently counted n=6 for two highly evolved species in sect. 30. 
Spachium, H. setosum and H. cumulicola. Although the numbers n = 12 and n=8 have each 
been recorded several times in this section, no intermediate ones have yet been found 
(2n = 18 for H. boreale (Kapoor, 1972), if correct, is likely to be a variant of n = 8, not a true 
intermediate.) The species with n=6 are not related to those with n= 8, however, but to H. 
denticulatum (n = 12, 24), of which they appear to be derivatives. Apart from the presence of 
hairs in H. setosum, the variation in the line H. denticulatum—®~H. setosum—+H. 
cumulicola is almost continuous, so that there is little ‘room’ for extinct forms with inter- 
mediate numbers. On the contrary, the change from 12 to 6 appears to have been sudden. 
How this change occurred remains to be discovered; possibly some form of Robertsonian 
(centric) fusion is involved. 


(b) Chromosome morphology 


Chromosomes of Hypericum are small (0°5-2°2 um long, cf. Robson & Adams (1968), 
Reynaud (1973, 1975)) and are often difficult to count satisfactorily. For these reasons no 
studies of chromosome morphology in the genus have been published. Variations in size 
within a complement do occur, however (cf. Robson & Adams, 1968), and acrocentric 
chromosomes with satellites have been depicted in H. kamtschaticum var. senanense 
(Matsuura & Sito, 1935). Therefore, although meta- and submetacentric centromeres seem 
to predominate in the genus, detailed comparative cytological studies might well yield 
information of evolutionary significance. 


| 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 167 


2n Authority Source Location of 
voucher 
R.& A. 
Webb (1980) U.S.A. (Alabama, New York, TENN 
R. & A. Pennsylvania, S. Carolina) 
Webb (1980) U.S.A. (New York) TENN 
18* Kapoor (1972) Canada (Nova Scotia) SMUH 
R. & A. 
16 I. & O. Hedberg (1977) Kenya (Mt. Elgon) UPS 
R. & A. 
14 R. & A. 


(c) Chromosome irregularities 


In the North American Hypericum punctatum (sect. 9. Hypericum), Hoar (1931) discovered 
structural hybridity (a ring of 16 chromosomes at meiosis); and Adams (1962a; Robson & 
Adams, 1968) recorded the same phenomenon in the closely related H. mitchellianum. This 
taxon is morphologically intermediate between H. graveolens and the form of H. punctatum 
that grows in the same area, and it hybridises with H. graveolens in nature (Culwell, 1970). It 
may therefore have the parentage H. graveolens x punctatum. Culwell, however, came to the 
conclusion that it was a good species. Adams (in Robson & Adams, 1968) counted n= 16 
doubtfully for H. graveolens but, according to Culwell (1970), now agrees that this count was 
an error for n = 8; whilst Bell (1965) counted n =7 for H. punctatum without mentioning any 
ring-formation. The relationships among these three taxa thus appear rather complex and 
would repay further investigation (see below, p. 168). 


Hybridisation in Hypericum 
(a) Natural hybrids 


Hypericum is not rich in natural hybrids; at least, not many are known. Where inter- 
gradation occurs, as it frequently does, it seems from morphological and distributional 
evidence to be most. often due to incomplete speciation. Only two taxa have been shown 
experimentally to be natural hybrids, but circumstantial evidence for the hybrid origin of 
some others is strong. 

In sect. 9. Hypericum, H. perforatum is tetraploid (n= 16) and shows signs of a hybrid 
origin. Thus the pollen is sometimes highly sterile (Noack, 1939) with lagging chromosomes 
at meiosis (Nielsen, 1924), and the embryo sacs are largely (92%) pseudogamous.* They 
develop either parthenogenetically (2n = 4x = 32) or rarely after fertilisation (2n = 6x = 48, as 
the pollen always undergoes normal meiosis, n = 16). Crosses between these forms result in 
pentaploid plants (2n = 40). This cytological variability is associated with considerable poly- 
morphism throughout the wide distributional range of H. perforatum, but the two types of 
variation have not yet been completely associated. Thus there is an apparent complete range 
of morphological variation from the typical broad-leaved, large-flowered form to the 
narrow-leaved, smaller-flowered var. angustifolium DC., in which the chromosomes are 
constantly reported as 2n=32 (except for the aberrant n=12 reported by Gagnieu & 
Wilhelm, 1965). The hexaploids are rare and do not seem to be morphologically distinct. On 


*Myers (1963) found a smaller proportion of pseudogamous embryo sacs in some American populations of H. 
perforatum. 


168 N. K. B. ROBSON 


the other hand, the count of 2n=16 for var. microphyllum DC. (= H. veronense) suggests 
that, as Schwarz (1965) concluded, this taxon may not belong to H. perforatum after all. 

From the above date it seems that H. perforatum may have resulted from an ancient 
hybridisation between two diploids with subsequent chromosome doubling. If this is indeed 
its origin, then one parent is almost certainly H. maculatum subsp. maculatum (n=8), 
which is distributed from central Siberia to the Pyrenees. On morphological and geographi- 
cal grounds the other parent could well be H. attenuatum Choisy. Although its chromosome 
number has not yet been counted, this species has the characters of H. perforatum that are 
absent from H. maculatum, and its distribution (in north-east Asia) overlaps that of the 
latter. 

Owing to its pseudogamously produced ovules (n=4x) and normally reduced pollen 
(n = 2x), the tetraploid H. perforatum produces triploid hybrids with diploids (H. maculatum 
subsp. maculatum, H. tetrapterum) as ovule parent and pentaploids when they are the pollen 
parent (Noack, 1939, 1941). With the autotetraploid H. maculatum subsp. obtusiusculum 
(cf. Robson 1957, 1958a, 19725), it usually produces tetraploid hybrids, which back-cross 
easily to either parent; but, on the rare occasions when an aposporous embryo sac of H. 
perforatum is fertilised, a hexaploid hybrid can occur (Fig. 55). The tetraploid F, hybrids (H. 
x desetangsii Lamotte nm. desetangsii) are intermediate between the parents; but back- 
crossing results in a continuous morphological series between them. On the other hand, 
whereas the triploid form of H. maculatum subsp. maculatum x perforatum (H. x desetangsii 
nm. carianthiacum (Frohl.) N. Robson) is recognisably intermediate (Fréhlich, 1960), the 
pentaploid form (H. x desetangsii nm. perforatiforme (Frohl.) N. Robson*) is almost indis- 
tinguishable from H. perforatum. 

As has already been mentioned (p. 167), sect. Hypericum also contains two structural 
hybrids, the N. American H. punctatum and H. mitchellianum. The latter hybridises in 


H. attenuatum 
(2n= 16 ?) 


H. maculatum ? 
subsp. maculatum 9 
(2n=16) > 
H. elegans 
H.maculatum (2n-32) 


subsp.obtusiusculum 


(2n=32) [sterile hybrid |@n-1e) 


H. triquetrifolium 
(2n = 16) 


H.x desetangsii 


nm. perforatiforme H.veronense 


(2n= 40) (2n-16) 
H.undulatum fH. tetrapterum 
(2n-16, 32) 
H. perforatum — H.perforatum 
(2n-382) (2n=48) 


H.x desetangsii 

7 nm. desetangsii 

H.x laschii (2n=32, 40,48) 
(2n=16) 


H, x medium 
(2n-24,40) 


Fig.55 Suggested relationships in the Hypericum perforatum group. 


*H. x desetangsii Lamotte nm. perforatiforme (Frohl.) N. Robson, comb. et stat. noy., Basionym: H. x carinthiacum 
a perforatiforme Frohl. in Mitt. Nat. Ver. Steiermark 51 : 229 (1915). 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 169 


nature with H. graveolens (Culwell, 1970) and is morphologically intermediate between H. 
punctatum and H. graveolens. Culwell crossed these species, but did not continue observing 
the resultant (healthy) seedlings until they reached maturity. 

In sect. 5. Androsaemum, the hybrid H. x inodorum Miller (H. hircinum x androsaemum) 
occurs in France and Spain, where the distributions of its parents overlap. It is intermediate 
in form between them. In gardens, where it appears to have arisen independently more than 
once in the 18th century (and later ?), it is much more variable, because the smaller forms of 
H. hircinum from various parts of the Mediterranean region (especially from the Balearic 
Islands) have crossed with different forms of the variable H. androsaemum. There is no 
apparent cytological barrier; both species have 2n=4x =40 and there is no evidence of 
reduced fertility in the hybrids (cf. Robson, 1975). 

In sect. 14. Oligostema, H. humifusum, H. linarifolium and H. australe form a ‘super- 
species’ in which intermediates occur where the component species overlap in distribution. 
Although it is not yet clear whether the intermediates are due to hybridisation or incomplete 
separation, Druce gave intermediates between H. humifusum and H. linarifolium from the 
Channel Islands the name H. x caesariense without validating it (cf. Robson, 1975). There 
appears to be a series of forms linking these species, both of which are diploids with 2n = 16 
chromosomes. On the other hand, the record from Germany of the intersectional hybrid H. 
humifusum x perforatum (H. x assurgens Peterm., nomen) (Kuntze, 1880) is based on a 
misidentification. The specimen appears to be a depauperate example of H. perforatum. 
Such a ‘wide’ cross would be unlikely to survive in a natural habitat. 

In sect. 20. Myriandra, where the known chromosome numbers are all n=9, Rehder 
(1910) reported three alleged hybrids growing spontaneously in the Arnold Arboretum, 
Mass., U.S.A., which he named respectively H. x arnoldianum (H. lobocarpum x galioides), 
H. x dawsonianum (H. lobocarpum x prolificum) and H. x nothum (H. kalmianum x densi- 
florum). After studying Rehder’s specimens, however, Adams (1972) concluded that only H. 
x dawsonianum had been assigned to the correct parentage. The plants of the latter and of H. 
x arnoldianum were fertile and showed no cytological irregularity or pollen sterility (Hoar & 
Haertl, 1932). On the bases of their intermediate morphology, Adams agreed that the 
parentage of H. x dawsonianum appeared to be H. lobocarpum x prolificum, but thought that 
that of H. x arnoldianum was H. densiflorum x lobocarpum, not the more distantly related 
H. galioides x lobocarpum. Adams decided that Rehder’s specimens of H. x nothum were 
assignable to H. densiflorum itself, not to H. densiflorum x kalmianum. Likewise, he 
concluded that H. x vanfleetii, a name which Rehder (1940) applied to a horticultural form 
intermediate between H. prolificum and H. frondosum, was not based on hybrid material but 
on an extreme form of H. prolificum. It is quite likely, however, that these last-named 
species, as well as other members of sect. 20. Myriandra not discussed above, can and do 
hybridise in cultivation; but, according to Adams, there appears to be very little interspecific 
hybridisation in nature among the woody members of this section. 

Finally, in sect. 30. Spachium natural hybridisation has been reported among some of the 
herbaceous species. The evidence is circumstantial; but it seems likely on morphological and 
distributional grounds that (i) H. x dissimulatum Bickn. (from Massachusetts, U.S.A.) is the 
result of crosses between H. canadense and H. mutilum or H. boreale (cf. Fernald, 1947 : 87; 
Webb, 1980*) and (ii) that H. gramineum and H. japonicum hybridise where their areas of 
distribution overlap in Bhutan (cf. Robson, 19734). 


(b) Artificial hybrids 


These have been confined mainly to two sections: 3. Ascyreia, where they have been 
produced for horticultural purposes, and 9. Hypericum, where their genetics, breeding 


*Webb (1980), on the bases of morphological intermediacy and reduced pollen stainability, recorded the following 
hybrids: H. boreale x canadense, H. boreale? x majus, H. canadense x gymnanthum?, H. canadense x majus, H. 
canadense x mutilum, H. majus x mutilum. He regards H. x dissimulatum as including recurrent hybrids of both H. 
canadense x mutilum and H. boreale x canadense. 


170 N. K. B. ROBSON 


systems and morphology have been studied. In both sections, hybrids that are the result of 
wide crosses are likely to lack chlorophyll to some extent, at least in the seedling stage. 


(i) Sect. Ascyreia. 

(a) H. x moseranum Luquet ex André (1889). This plant was the result of the first recorded 
hybridisation in sect. Ascyreia, made by Moser, a nurseryman at Versailles. He crossed 
Hypericum patulum ¢ (introduced to Europe from Japan by Oldham in 1862) with H. 
calycinum L. &. The resultant hybrid was intermediate between the parents and 
apparently fertile. It has become popular in gardens. Considering the parental chromosome 
numbers (respectively 2n = 36, 20), it would be expected to have 2n = 28; but Sugiura (1944) 
recorded n = 18, whereas Dr Mary Gibby found 2n =c. 50 (see Table 7, p. 152). The former 
count may have been based on a misidentification, but the latter suggests some cytological 
instability in the hybrid. Genetical instability no doubt produced the sport with variegated 
(white- to pink-margined leaves) which Maumené named var. tricolor and Rehder reduced 
to a forma. The correct category, of course, is nothomorph.* The leaves and flowers are 
smaller and appear less healthy than those of nm. moseranum. 

(8) H. x ‘Hidcote’. When this plant was first sent to a Royal Horticultural Society’s show, 
where it was exhibited by Hilling’s Nursery, Chobham, Surrey, it was said to have been 
possibly introduced from Yunnan by Lawrence Johnston, the owner of Hidcote Manor, 
Gloucestershire (Synge, 1950). Other suggested sources of introduction have been (i) the 
famous collector of Himalayan plants George Forrest and (ii) an East African garden (cf. 
Fletcher, 1955). I have, however, seen no specimens from a wild source. An origin by 
hybridisation at Hidcote seems much more likely (cf. Robson, 1970). This possibility was 
also mentioned in 1950. The records at Hidcote Manor of the origin of their plants have 
unfortunately been lost (Head Gardener, Hidcote Manor, 1961, pers. comm.). On the other 
hand, Stern (1960) stated that his Hypericum x ‘Hidcote’ plants were produced from cuttings 
received from W. Miller-Christy of Watergate, Chichester, Sussex in the 1920s. At any rate, 
H. x ‘Hidcote’ behaves like a hybrid. The pollen is mostly sterile (cf. Thomas, 1970; also P. 
Dummer, pers. comm.), with lagging univalents at meiosis of the pollen-mother-cells, and 
seed is never set, the whole flower falling after the petals and stamens have been shed. Ifit isa 
hybrid, then the relatively densely reticulate leaf-venation, large sepals and long styles, 
together with its hardiness, suggest that one parent is probably H. calycinum. The identity of 
the other parent, however, is still in doubt. 

At first I thought that H. forrestii could be the missing parent (Robson, 1970) and 
encouraged two hybridisers, Mr Donald Walker, North Mymms, Hatfield, Herts. and Mr 
Peter Dummer of Hilliers’ Nurseries, Winchester, Hants, to make the cross H. forrestii x 
calycinum. Both hybridisers produced plants that incorporate certain of the characters of H. 
x ‘Hidcote’, some of which look superficially like it; but neither set of progeny contained an 
exact match for it. In particular, although its deep orange anthers appeared in some of the 
seedlings, in none were the stamen filaments as short as they are in H. x ‘Hidcote’. It would 
seem, then, that the other parent of the latter is likely to be a short-stamened species. 

On a recent visit to Hidcote Manor, I found that the plant known to me as H. beanii ‘Gold 
Cup’ was being grown there under the name H. ‘Lawrence Johnston’. The only other 
members of sect. 3. Ascyreia in the garden were H. calycinum and H. x ‘Hidcote’. It then 
occurred to me that H. beanii ‘Gold Cup’ was morphologically suitable to have been the 
other parent of H. x ‘Hidcote’. It has the narrower, more acute leaves and the shorter stamens 
that that parent must have had, and its other characters are also suitable. From plants of H. 
“Lawrence Johnston’ (obtained through the kindness of Mr P. Nichols, the Head Gardener at 
Hidcote Manor), Dr Gibby was able to make a chromosome count of 2n = 30, thus establish- 
ing that this taxon is fit cytologically to be a parent of H. x ‘Hidcote’ (see p. 151). 


*H. x moseranum nm. tricolor (Maumené) N. Robson, stat. nov. 
H. x moseranum vat. tricolor Maumené in Le Jardin 8 : 186 (1894). 
H. x moseranum forma tricolor (Maumené) Rehder, Man. Cult. Trees & Shrubs : 463 (1949). 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 171 


From the above considerations, then, it seems very likely that H. x ‘Hidcote’ arose (or was 
created ?) at Hidcote Manor as a hybrid between H. calycinum and H. beanii ‘Gold Cup’. 
The true status of the latter taxon remains to be determined. 

H. x “Hidcote’, like H. x moseranum, has given rise to a variegated form. As they age, most 
plants of this hybrid produce shoots bearing narrow variegated leaves, a phenomenon which 
has been interpreted by several horticulturalists as indicating genetic deterioration of the 
clone (cf. Bean, 1973) or a virus attack. Propagation from these shoots results in weakly 
plants with small flowers comparable with those of H. x moseranum nm. tricolor, but more 
erect and with no pink tinges in the white areas of the leaf. 


(y) Other hybrids. The only other hybrid in sect. 3. Ascyreia that is known to me is 
Hypericum kouytchense x calycinum, also made by Mr Donald Walker. He reports that the 
petals of the hybrid open tardily or even remain imprisoned in the calyx, and the capsules do 
not turn red like those of H. kouytchense. The flowers, however are larger than in that 
species, and the hybrid set apparently ripe seed. 


(ii) Sect. Androsaemum. The parentage of H. x indorum Miller (H. hircinum x 
androsaemum) was deduced from circumstantial evidence, viz. intermediate morphology, 
occurrence in nature within the common distributional area of both parents, and the corre- 
spondence of various cultivated forms to variations in the parents (Robson, 19735). Artificial 
crosses of H. hircinum and H. androsaemum have, however, also yielded plants indistin- 
guishable from H. x inodorum (G. Thomas, 1971, pers. comm., D. Walker, 1973, pers. 
comm.). Mr Walker has been growing seedlings resulting from selfing H. x inodorum in order 
to study the segregation, but his results are not yet to hand. 


(iii) Sect. Hypericum and intersectional hybrids. 

(a) Hybrids not involving H. perforatum. The discovery of variegation in some experi- 
mentally produced Hypericum hybrids (Fahrenholtz, 1927) led to extensive hybridisation 
studies in the genus by Noack (1930-1941) and Herbst (1935). These were concerned mainly 
with H. tetrapterum (‘H acutum’), H. maculatum subsp. maculatum (‘H. quadrangulum’) 
and H. perforatum, all in sect. 9. Hypericum, but also involved wider crosses with H. 
calycinum (sect. 3. Ascyreia), H. olympicum (‘H. polyphyllum’) (sect. 10. Olympia), H. 
rumeliacum (sect. 13. Drosocarpium), H. pulchrum and H. hirsutum (sect. 18. Taenio- 
carpium) and H. montanum (sect. 27. Adenosepalum). Excluding those involving H. 
perforatum, which require special consideration, the results of these hybridisations may be 
summarised as follows (Fig. 56): 

(1) Crosses involving different chromosome numbers either set no seed or occasionally 
resulted in seedlings only. 

(2) Crosses involving the same chromosome number sometimes set seed. Intrasectional 
ones succeeded (sect. 9. Hypericum) or not (sect. 18. Taeniocarpium),; but the only success- 
ful ‘wide’ crosses involved H. montanum and H. tetrapterum. In all these crosses the fertility 
was more or less low, and reciprocal ones sometimes behaved differently. This was particu- 
larly noticeable with regard to the type, distribution and persistence of variegation in the 
hybrids, a circumstance which led to controversy (cf. Correns, 1931). Herbst (1935), 
following his teacher Renner, interpreted the various variegations in terms of the ability of 
chloroplasts transferred to the ovule parent via the pollen tube to develop chlorophyll in 
alien cytoplasm. Noack (e.g. 1937a, b), however, denied that all the variations could be 
explained thus. The matter seems never to have been resolved satisfactorily. 

Variegation is not always a sign of hybridity in Hypericum. As Herbst pointed out, it may 
be due to growth in an unfavourable habitat; and it has also been found in natural conditions 
which did not appear unfavourable, e.g. in H. elegans (sect. 9. Hypericum) (Pringsheim & 
Schwarz, 1933). 


(8) H. perforatum and its hybrids. Herbst (1935) was cautious in interpreting his 
hybridisations involving Hypericum perforatum, suspecting some irregularity in the 
breeding system, and Noack (1939, 1941) confirmed his suspicions. 


172 N. K. B. ROBSON 
Sect.9 Hypericum 


H. maculatum (16) 
_SSp. maculatum 


H. maculatum (32) 
ssp. obtusiusculum 


Sect. 3 Ascyreia 


Z 4 
eS a pis ae < le Gea a era R= = — — X= H. perforatum (32) 
S r 
| \ 
Sect. 27 | 
Adenosepajlum 5n 
| only 
H. montanum 
| (6) 
Sect. 18 


Taeniocarpium 


\ H.pulchrum (1g) 
Viney 


sees H. hirsutum (18) 


fF? 
7 


fully fertile 
eos, Feduced fertility ~ iS 


(to flower or fruit) Mork ress cae . a 
d fertilit , 
Geedlings only)” H. rumeliacum (14) ----—- H.olympicum (18) 
renee a Sect.13 Drosocarpium Sect.10 Olympia 


weary CrOSS Not made 


Fig. 56 Summary of experimental crosses carried out by Fahrenholtz (1927), Noack (1930-41) 
and Herbst (1935). The nomenclature used by these authors has been amended. 


H. perforatum proved to have a largely apomictic form of reproduction, but to require 
pollination before seeds would set, 1.e. it is pseudogamous. This is due to the almost invari- 
able need to fertilise the endosperm nucleus before the embryo can develop. Both sexual and 
aposporous embryos have the same chromosome number (2n = 32), so that the only means of 
determining the method of reproduction is by counting the chromosome number of the 
endosperm nucleus: 16 + 16+ 16=48 in a sexually produced seed, 32 +32 + 16=80 in an 
aposporous one. 

Since H. perforatum is tetraploid, its aposporous embryo sacs are tetraploid and its pollen 
diploid. Therefore as mentioned above (p. 168), in crosses with diploids with n=8 or 9 the 
number of the hybrid differs according to whether H. perforatum is the pollen parent 
(2n=3x=16+8 or 9 =24 or 25) or the ovule parent (2n = 5x = 32 +8 or 9 =40 or 41). Since 
H. perforatum contributes respectively twice and four times as many chromosomes as does 
the diploid parent, it is not surprising that Noack found the hybrids to be more similar 
morphologically to H. perforatum or, in the case of some of the pentaploids, indistinguish- 
able from it (cf. Fig. 55). 

Neither Noack nor Herbst was aware that H. maculatum includes plants with two different 
chromosome numbers (n=8, 16), those with n= 16 being autotetraploid (Robson, 1956, 
1957). Spontaneous hybrids between the latter (subsp. obtusiusculum (Tourlet) Hayek) and 
H. perforatum are common, the progeny being apparently fuliy fertile (H. x desetangsii nm. 
desetangsii), and this hybrid has also been produced experimentally (cf. Robson, 1975). 


(y) Other artificial hybrids. Myers (1963) crossed H. perforatum and H. punctatum 
(sect. 9. Hypericum), H. frondosum (diploid and colchicine-induced tetraploid) and H. 
prolificum (sect. 20. Myriandra), H. floribundum (= H. canariense, sect. 21. Webbia), and 
Triadenum virginicum in all combinations. He obtained seedlings from the H. frondosum x 
prolificum cross (possibly in both directions), but not from the other crosses. 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 173 


10. Distribution 


Introduction 

It is a truism to state that meaningful phytogeographic studies can be made only on mono- 
phyletic taxa; and it was partly for this reason that Hypericum (a clearly monophyletic 
genus) was chosen for study. It was also for this reason that I have endeavoured to discover 
the evolutionary trends in the genus as a whole, as well as in each section of it. For, even 
though recognition of some of the sections (e.g. 22. Arthrophyllum, 28. Elodes) has resulted 
in paraphyly, these sections themselves are monophyletic with regard to their immediate 
ancestors (for the above sections respectively 21. Webbia and 27. Adenosepalum). The only 
exception would appear to be sect. 9. Hypericum, where it now appears that H. concinnum 
Bentham (endemic to the mountains of central California) is related to a different part of 
sect. 7. Roscyna from that which gave rise to the rest of the section. It is therefore necessary 
to place this species in a monotypic section.* 

It is also true that, whilst merely plotting the distribution of a monophyletic taxon may 
produce a pattern that is suitable for phytogeographical analysis, the introduction of cladistic 
considerations (e.g. arrows indicating morphological trends) will make the distribution map 
more meaningful. Although it may seem self-evident that a geographical replacement series 
of taxa showing correlated primitive to derived characters gives an indication of the direction 
of evolutionary development, some authors have taken a contrary view (see Ball, 1976). 
They argue that the derived forms force the ancestral ones to peripheral areas. Ball 
(1976 : 420) himself finds it difficult to decide a priori between these two possibilities, as 
indeed it may be in an isolated instance. However, where repeated geographical trends form 
a logical pattern in a monophyletic taxon, as they do in Hypericum, it seems safe to assume 
that they do indicate evolutionary progressions; and where these patterns are interrupted by 
wide areas of ocean, they can throw light on aspects of the geological history of the earth. In 
such cases, Hennig’s Phylogenetic Intermediate and Multiple Sister-group Rules and 
Ashlock’s Drift Sequence Rule may be applied (Ashlock, 1974). 

There is no essential difference between a geographical trend with gaps in distribution 
(possibly resulting in speciation) and one in which the changes take place gradually over a 
continuous area within a single species (although, here, ‘continuous’ is a relative term). The 
continuous area must have resulted from gradual colonisation (dispersal) and the gaps from 
disruptive processes such as extinction, mountain-formation and continental rifting 
(producing vicariance). If a third process, namely long-distance dispersal, is involved, then a 
disruption of the logical pattern mentioned above is to be expected. Thus, the (mainly 
zoological) argument between the proponents of ‘vicariance’ and ‘dispersal’ (cf. Croizat, 
Nelson & Rosen, 1974; Ball, 1976) would appear to be the old one of the role of long- 
distance dispersal in explaining wide disjunctions in distribution, as the other two ‘theories’ 
(dispersal and vicariance) really concern two aspects of the same phenomenon. Since gradual 
dispersal is the normal means of area extension, then ‘Occam’s Razor’ would preclude the 
invocation of a hypothesis of long-distance dispersal unless a given distribution is not 
explicable in terms of normal dispersal. The proponents of the ‘vicariance’ theory (e.g. 
Nelson, 1976) are at pains to contrast it with ideas of ‘centre of origin’, suggesting that 
apparent centres from which related taxa (e.g. species in a genus) radiate result from vicari- 
ance rather than from gradual dispersal. If, however, dispersal and evolutionary diversifi- 
cation from a localised area are followed by isolation of populations, the concept of ‘centre of 
origin’ seems to be quite valid. The gradual dissection of a distributional area with time often 
results in the most primitive taxa in an evolutionary progression having a reduced, relict 


*Hypericum sect. 9a. Concinna, sect. nov. Sectioni 9. Hypericum affinis, sed foliis concoloris, saepe conduplicatis vel 
falcatis, interdum nigropunctatis; sepalis magnis imbricatis inaequalis, interdum nigropunctatis; antheris connectivo 
succinoglanduloso; differt. 

Type: H. concinnum Bentham. 

DIsTRIBUTION: California. 

1 species. 

Basic chromosome number: 8; ploidy 2x. 


ss 
. ‘ “(O861 “souos) . 
SNAADIOAYIOC “JOOS SNdADIOAYIO SOPNIIUI DaUWUD “‘DIuldADsO] dag pue wnjvdjadi4 J wiusuniDYy dpluag “DIUIIADSDAD *\O9S SNdADIOAYIO “DIpaayy sopnjoul DIUIDIDD, 


EE EE eee 


x x x x x x x x x wno14add Fy 
DISDUOJUDS 


DAUAOY J, 
winuapDd1d J 
xX winjAXoJDAD 

Da” 
DUDSUNAD TT 
WNUAAdSOAOS J 
x DIUS1A 
avaplooodAyYy 
(p) 
e19U93 SUIUIBUIOY 
xX BIBUIS 1940 p 
2} DISHID 
x DIUIIADL) 
DINIDI QUO] Y 
wunisunjoydnraq 
aeaprloisn{a 
(p) 
e19U93 19Y1O 
xX piuoydudgy 
x DAAIZNOAJUO PV 

DUSapoluad 
IeIPlOaqouo1Opy 
% uodnaUuoj1da0g 

DIunAgaT 
Dilusapopuy 
x xX ~DIUWD 
DIWWMDUDIDG 
x x unjjcy dojo) 
pnsayw 
avaproy[Aydoyea 
xX (8) 
b19U93 INYO 
DLLADIN 
pyauuog 
xX WNIWDIO| 
svaplonauuog 


xx Kx 


~*~ xX 
mK xK x 


xx 
Kee eG pe pe 
* 


N. K. B. ROBSON 
xx 
x KK 
x KK »* 


~ x 
xx 


BIsejelisny  eiuopalea purleoZ eiskeyeyyY =eyuey souareoseyy RIV BWV N eBoWoWy eoloUYy 
JO SOY MON AVON 79. % iets) # yInos 2% vISV Yinos [e1jUus) 
eYyelsny eIsy q'S Weipuy sIeosesepeypy 9 dol dura} ® YLION 

‘odoin 


a 


SeIQINH 941 JO uornquIsIq g 2I1qBL 


174 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) GS 


area, those immediately derived having several isolated areas, and the more advanced having 
larger, more continuous areas (cf. Fig. 7, p. 72; Kubitzki, 1977). These three types of taxa 
may occupy areas in a linear sequence, in which case the ‘centre of origin’ is at one end of the 
sequence; or they may occupy concentric areas, in which case the centre of origin is a true 
centre. In both cases, dispersal and vicariance are involved. 

In Hypericum there are many examples of long-distance vicariance, some of which are 
trans-oceanic; and they cannot all be explained by the same hypothesis, whether it be 
gradual dispersal and vicariance or long-distance dispersal. 


Distribution of the subfamilies of the Guttiferae 
(a) Moronobeoideae and Bonnetioideae 


The subfamilies of the Guttiferae show vicariance to varying extents, some of their distri- 
butions indicating great age (Table 8). This can be said with confidence in some groups (e.g. 
the Moronobeoideae—tropical America, tropical Africa and Madagascar, New Caledonia), 
owing to the size and weight of their fruits; whilst, even where the seeds are small and 
partially winged, there is no evidence that they are adapted to, or have undergone long- 
distance dispersal (e.g. the Bonnetioideae—tropical America, Malaysia). Apart from the 
Hypericoideae, the other subfamilies (Calophylloideae and Clusioideae) are both pan- 
tropical and more-or-less heterogeneous and will not be considered further here. 

In the Moronobeoideae (Fig. 57) the most primitive genera (Pentadesma, Montrouziera) 
are the farthest apart spatially (Africa: Guinea-Congo and New Caledonia, respectively), and 
one of the most advanced (Symphonia) has one species which occurs widely in the tropics of 
both sides of the Atlantic and several derived species in Madagascar. The intermediate 
genera are all tropical American. There thus appears to have been an initial east-west 
divergence that isolated Pentadesma and Montrouziera from each other and from the rest of 
the subfamily, followed by evolution to generic level in tropical America (5 genera) with one 
of these genera (Symphonia) spreading back eastwards to Africa and Madagascar. It 
speciated in Madagascar, but not elsewhere. 

In the Bonnetioideae (Fig. 58) the trends are less clear, because all the genera except one 
are in tropical America. As we have seen, some of these (e.g. Neblinaria) are relicts which 
retain some very primitive characters such as the virtual lack of a leaf midrib. In other 
respects, however, the most primitive genus is the isolated one, Ploiarium, which has one 
species (P. alternifolium Vahl) in southern Thailand, Malaya, northern Sumatra and 
northern Borneo, and a second (P. sessilis Scheffer) in extreme-western New Guinea. It is 
possible to see these two groups—the American and the Malaysian—as the result of an 
ancient east-west divergence; but there is not enough evidence to indicate the area from 
which this divergence occurred. 


(b) Hypericoideae 


It was suggested above (p. 64) that the three tribes of the Hypericoideae were respectively 
related to different parts of the Bonnetioideae, and their distributions and trends bridge the 
gap in the latter’s distribution. 

The Vismieae, which are closest to the mainly Amazonian genus Caraipa, is itself 
primitively Amazonian in Vismia. Thence this genus extends to nearly all parts of tropical 
_ America (except the West Indies north of Tobago) and into tropical Africa, where there 
_ occur two groups of species each related to a different American species (Robson, ined.). 
‘| Each group has given rise to a genus with modified fruit; in one (Harungana), each of the five 
multi-seeded loculi has become a pyrene (i.e. has a hard endocarp), whereas in the other 
(Psorospermum) each loculus has become one-seeded. Both genera occur in Madagascar, 
and Harungana extends into the Mascarenes (Mauritius); but Harungana has scarcely 
speciated (two species), whereas Psorospermum has five species on the African mainland 
(Bamps, 1966) and a considerably larger number in Madagascar, all or mostly epibiotic. 


1 
x Ps | 
fs ati eee “<4 
\ NO i 
\ Sg 1 
\ ‘oe 
\ Nassau amos | 
NS H Se 8 H 
\ Iso ee 
\ Is i 
N\ 1 aot | Sa 
ARG =e ° ae = 
xX be WT =8Pa==S5l IT 
& a 
N | -~ 
~ | ~~ 
\ | | 
> | | | i 
N | « H ~ 
Se e | | ne 
=. | = erry | ~ 

Sali ° ! | 

ws | | 

| 


N. K. B. ROBSON 


ROYAL BOTANIC GARDENS . KEW 


of,» See 
/ 

/ 

/ 

von 


| 
| | 
aa 
I ~ :\ 
| the NTN 
H i Soy \ 
| eae Se 
Sse Ss 5 \ al \ 
i —af ae ee eae de | Get > ee Soe + | 2 \ 
t | (i Die Se alee GaSe ane pri alan eT re i 
ae | i | ee! 
2 ee S| | | ae 
— “Se ees re oe ac OR Ear. aaa tC gee eat = — a a aS 


~ 10/64 


C.B.H. 4623 - Wt. 34154 - Dd. 10 - 


other genera. 


Fig.57 Distribution of the Moronobeoideae = = Symphonia, 


We 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 


° 
= Bi 
Al S fe) 
mo ~ = 
cme ae = Y 
Be o° | Se =~ 
° cA ° i aS se) 
oe eS ~ oe 
Q7 oid oa 
- ~S 
- mK ~ 


| 

/ 
ee 

/ 
/ 
/ 
/ 
—7/60° 


/ 
4 
MOLLWEIDE S 


se 


\ , 
\ ; 
\ 
ae 
i] 
yh 
/ 
> 
/ 
/ 
/ 
/ 


woe} 
- 


the 


N. K. B. ROBSON 


178 


EQUATOR 


ind 
wi 
= 
Ww 
as 
ee 
at 
L 
o 
2 
2 
o 
Z 
fe) 
wn 
[4 
we 
m 
2 
> 
~ 
a 
Z 
= 
< 
m4 
[a} 
w 
4 


= 
re 
x 
4 
z 
w 
Q 
4 
= 
6 
zg 
z 
= 
3 
a 
~ 
= 
S) 
« 


- 10/64 
Fig. 58 Distribution of the Bonnetioideae: :3- 


C.B.H. 4623 - Wt. 34154 - Dd. 10- 


-:. Bonnetioid genera, = = =Caraipoid genera, 


Kielmeyera. 


rire Xr) 


179 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 


ee 
A UR-n 
aay 
eo $ % 
oa So)  BeSea= 6) 
2 -_- - 
aera 
os ee 
A 2a H 
7 oa | 
| bee | 
Me ls | 
7 Ae | 
/ c & 
/ Pe it fe —— =) 
/ 7 ae aad =| 
ve 7 - 


-------}------ 


\ 
\ 


\ 


\ 
Neen 
\ 
\ 


740 
/ 
PROJECTION 


/ 
/ 
/ 
—7/60° 


4 
Ce 
MOLLWEIDE S 


= 
Eo 


th 
one aa 


oa ee 


——------+ ------+---- 


SS 
SS 


------ 


| 
S 
& AF 
fe) 
[4 
ee] 
vy | 
: | 
a Be 
LS Fem ff anc ee = = 
| ee 
1 at ! 
eu 
\ I ne 
i | =f -—---~ 
| 
\ K | | 
4 1 Pa 1 Y H | 
a See | 
ON \\ ss | 
aaa 2 ee | 
XX NN " 5 y ES SSaass Bees ei ge aaah 3 
NS Io. | ai 
ee |! 
NY dose | | 
° 8 \ 
9 Si oS ? 
N oS “l ° 
Sct ~tp-4-------- 3 
: i ! 
or) Tease ae 
_ 0. a — - — — 
a % 


REDRAWN by M.GRIERSON after H.R.HEWER 


ROYAL BOTANIC GARDENS . KEW 


vi 

| 

| 

| 

| 

| 

om \ 

ies Ne 

| RS | iN 

| Se " BN 

| 
ls SS & | N FN 

=a Clea Vee 

(Geo AS Ne 

! sat ‘ie \ 

as ‘\ \ 

| | SS: \ 

| | See UI 

1 | SY \ 
Sr : \ 

r es Li 

I ail 
rs SRSA Sea [| ree oe 

S % = 


- 10/64 
Fig. 59 Distribution of the Hypericoideae—Vismieae and Cratoxyleae: 


C.B.H. 4623 - Wt. 34154 - Dd. 10 - 


Vismieae, 


Cratoxyleae. 


° 
AG 
= 


181 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 


\ 

\ 
\ 
\ 
\ 

= | 
/ 
| 

/ 


\ 


PROJECTION 


MOLLWEIDE S$ 


\ 


sete) go eae 


| 
-7T+ 
|, 
( 


Jl 
| 
! 


7 p= 
4 
/ 


7 


<a 
/ 
/ 
i 7, 
/ 
52SS5= 
/ 
/ 


Lesa = 
Awan 
——— 
ee ee 7 
ae x 
A 
BA 
7 oo 
7 
=--% --f~---- 


a 


182 N. K. B. ROBSON 


(Perrier de la Bathie (1951) describes 26 species, all of which are said to be endemic.) The 
trends in the Vismieae are thus clearly from the New World to the Old (Fig. 59). 

In the Cratoxyleae, the primary vicariance is between Eliea (Madagascar, monotypic) and 
Cratoxylum (Burma and south China to western Malesia, 6 species) (Gogelein, 1967; Robson 
1972a, 1974). Neither genus is clearly more primitive than the other (cf. Baas, 1970), 
although the range of variation is much greater in Cratoxylum. The other two genera, which 
both show trans-Pacific links, seem to have been derived independently from Cratoxylum. 
One of these, Thornea, comprises two species of relict habitats in the Mexico-Guatemala 
border area (Breedlove & McClintock, 1976); the other, Triadenum, has one species (T. 
breviflorum (Wall. ex Dyer) Y. Kimura) in Assam (Khasia) and Taiwan, one species (T. 
japonicum Makino) in Japan, Korea and adjacent U.S.S.R., and two pairs of species—or two 
species each with two subspecies—in eastern U.S.A. and Canada (Robson, 1972a: fig. 2, 
1976). Since the American species of Triadenum are, in general, more advanced than the 
Asian ones, an evolutionary trend eastward across the Pacific is apparent here also (Fig. 59). 

The Hypericeae, as we have seen, comprise two genera, Hypericum and Santomasia, each 
of which has primitive characters not shared with the other. Santomasia, like Thornea, is 
confined to the Mexico-Guatemala border region; but its links are trans-Atlantic, not trans- 
Pacific, in that the primitive species of Hypericum (H. bequaertii, sect. 1. Campylosporus) is 
in tropical Africa. The primary vicariance in the Hypericeae is thus trans-Atlantic (Robson, 
1977 : fig. 4, 1979 : table 2), and there is no clear directional indication from morphology. 

In the Hypericoideae, therefore, the Vismieae show two separate trans-Atlantic trends 
from west to east, each of which continues to Madagascar (and one to Mauritius); the 
Cratoxyleae show a trans-Indian-Ocean vicariance followed by two separate trans-Pacific 
trends, one tropical and the other temperate; and the Hypericeae show an initial trans- 
Atlantic vicariance. 


Distribution of Hypericum 
(a) Sect. 1. Campylosporus and its immediate derivatives 


Sect. 1. Campylosporus, which includes the most primitive species in Hypericum, is 
confined to tropical and southern Africa, south-western Arabia and nearby islands 
(Fernando Poo, Socotra, Grande Comore, Madagascar and Réunion) (Fig. 60; Robson, 
1977 : figs 2, 4). Being confined to higher altitudes, the species all have restricted or more-or- 
less discontinuous distributions along the east African mountain chain from Eritrea to 
northern Cape Province (Bamps, 1971 : maps 72-76; Robson, 1979 : maps 1, 2). Two (H. 
revolutum, H. roeperanum) occur also in the Cameroon highlands, and one of these (H. 
roeperanum) has a wider western distribution, being found in Angola (Huila), Nigeria 
(Bauchi plateau) and Guinée (Fouta Djalon). Whilst such distribution patterns are typical of 
Afro-montane species (cf. Morton, 1972), a closer examination of the distributions of the 
various members of this section reveals several points of interest. 

If the species of sect. Campylosporus are arranged in a relationship diagram (Figs 3, 61), it 
may be seen that the most primitive one (H. bequaertii) has a restricted (relict ?) distribution 
in the Ruwenzori Mountains. Its closest relative, H. revolutum subsp. keniense, also occurs 
on Ruwenzori but is found on the Virungas and the east African ‘inselbergs’ as well; and it, in 
turn, is incompletely distinct from the more advanced subsp. revolutum, which is widespread 
in Africa and occurs in Arabia. 

The other taxa that appear to be directly related to subsp. keniense are all widely disjunct 
with the exception of H. quartinianum, which has probably spread into the area of subsp. 
keniense from much further north. Within sect. Campylosporus the disjunctions are to 
Socotra (H. socotranum) and Réunion (H. lanceolatum subsp. angustifolium), the latter 
being particularly similar morphologically to H. revolutum subsp. keniense (Robson, 1979). 
In addition, the two sections either confined to Central and South America (29. Brathys) or 
with the primitive species there (30. Spachium) appear to have direct relationships with H. 
revolutum subsp. keniense. The taxa directly derived from H. bequaertii and H. revolutum 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 183 


subsp. keniense thus comprise two on the other side of the Atlantic Ocean, two on islands 

noted for the occurrence of relict taxa, and two on the African mainland. 

When these sections are associated with the relevant parts of sect. Campylosporus (Fig. 
61), it can be seen that, in general, the taxa ‘radiating’ from H. revolutum subsp. keniense (in 
terms of morphology and other ‘internal’ characteristics) are distributed round its area in a 
similar ‘radiating’ fashion; and there is a second ‘radiation’ from the north of the area of 
sect. Campylosporus from the Ethiopia-Socotra region (Fig. 60). 

The southern ‘radiation’ comprises five trends, which will be considered in an anti- 
clockwise direction round the area of H. revolutum subsp. keniense: 

(i) To the north-west, sect. 29. Brathys has its most primitive species (H. styphelioides) in 

Belize and Cuba. 

(ii) To the south-west, sect. 30. Spachium has its most primitive species (H. rigidum = H. 

meridionale L. B. Sm.) in south-eastern Brazil (centred in Parana). 

(i11) To the south-east is H. Janceolatum (in Réunion and Grande Comore) with a derivative 

species, H. madagascariense, in north and central Madagascar and a derivative section, 
26. Humifusoideum, of which the most primitive species (H. sewense) occurs in the 
Madang District of northern New Guinea (iii a). 

(iv) To the north-east, sect. 3. Ascyreia, which has its most primitive species (H. mysurense) 

in Sri Lanka and southern India, is in turn the source of the Eurasian and North American 

group that comprises sects 4-19. 

(v) To the north-north-east (in Socotra) is H. socotranum, in which, I have included the 
plants that previous authors (e.g. Hutchinson, 1959 : 297) regarded as belonging to H. 
mysurense. H. socotranum, which comprises three taxa (subspecies ?), is in turn the 
most closely related species in sect. Campylosporus to three sections. Sect. 23. Tri- 
adenioides has its most primitive species (H. scopulorum) in Socotra itself (v a), whereas 
the most primitive form of the basic species of sect. 25. Adenotrias (H. aegypticum) is in 
southern Morocco. The monotypic sect. 24. Heterophylla (H. heterophyllum) from 
north-eastern Turkey is more advanced than some forms of H. aegypticum in all 
characters except for those of the flower, which lacks the specialised pollination 
syndrome of sect. Adenotrias. These two sections have therefore been treated as parts of 
a single ‘radiation’ from H. socotranum (v b). 

The northern ‘radiation’ comprises the derivatives of two taxa, H. revolutum subsp. 
revolutum and H. quartinianum, both of which have quite wide distributions in Africa and 
also reach Arabia. They are not particularly closely related to each other (Fig. 61), but both 
appear to have given rise to distinct sections northward from the ‘centre’ of sect. Campylo- 
Sporus: 

(vi) H. revolutum subsp. revolutum is relatively constant in morphology throughout most of 
its wide range from Yemen south to Cape Province and west to Fernando Po, but in 
Ethiopia and Eritrea broader-leaved forms occur. The monotypic sect. 2. Psorophytum 
(H. balearicum), which is confined to the Balearic Islands (Fig. 70, p. 202), has 
apparently been derived from such broader-leaved forms by complete separation of the 
styles, reduction of the inner floral whorls to tetramery, and enlargement of the 
pellucid glands (vi a). For trend vi 6 see footnote on p. 187. 

(vii) H. quartinianum has a smaller range than H. revolutum, scarcely penetrating south- 
ward beyond Tanzania; but it seems to have been a fertile source of derivative groups. 
In this trend there is a tendency for the styles to become completely united or at least 
appressed; and in H. quartinianum and the more specialised H. roeperanum this 
condition is sometimes achieved. The latter species has the widest distribution in sect. 
Campylosporus, not only extending from Ethiopia to the Transvaal, but occurring 
disjunctly in Cameroon, Niger, Guinée and Angola. The complete union of styles is 
constant in H. synstylum, a local species of eastern Ethiopia and Somalia, which is the 
nearest relative of H. frondosum, the most primitive species in the north American sect. 
20. Myriandra (vii b). Here the styles are also completely united, at least during 
flowering. 


N. K. B. ROBSON 


184 


a a om 


a 
wi 
= 
iy 
25 
CA 
“16 
ra 
2 
‘ss 
Zz 
fe) 
7) 
[-4 
a 
a 
(©) 
= 


= 
i) 
=< 
uo 
z 
wy 
Qa 
« 
< 
ie) 
XZ 
ze 
< 
is 
So 
a 
= 
= 
f°} 
« 


REDRAWN by 


~ 10/64 
Fig. 60 Distribution of sect. 1. Campylosporus and of the most primitive members of sections 


C.B.H. 4623 - Wt. 34154 - Dd. 10 - 


directly related to it. Arabic numerals indicate sections; Roman numerals indicate trends (see 
Fig. 61 and footnote, p. 187).¢**«=area of distribution of H. revolutum subsp. keniense. x 


of H. bequaertii. 


area 


185 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 


i 
dy 
i 


5 2% 
ee aa ee ee | 
SS oe 2 z 
~ 
-T i ~~ os ie) 
4 ° as |i S36 5 
ee Sf} --—-—-—~__ : Ts ire 
—— = . = = = 
or ae oe | 9 
- | — a“ 
| ~ a 
| ~ 

| | | a) 
; ! : | | wu 
| ° | | a 
ae = 
w 
| Q, | | $ 
1S. soe SF oe oe ——— — if 2 
ae - Ky ae te 1 a 
= O°) = e) 
| = 

| 

| 

| 

| 

| 


a 
I 7 
re Sale 


186 N. K. B. ROBSON 


Sect.28 Elodes 
(w. Europe, Azores) 


Sect.20 Myriandra t , 
N.E.America to Guatemala, t 7, 
2 _ : 
CE Ee aes Sect.27 Adenosepalum we 


Bermuda) # 
i] 
i] 
i} 


H. synstylum 
(Ethiopia, Somalia) 


4 
/ 


Sect.21 Webbia 
(Canary Is., Madeira) 


A 


(Africa, Europe, S.W. Asia, 
India to China 


f 


U] 


t 
Wii) 7” 


= 


H.gnidiifolium 
(Ethiopia) 
vb) 


Sect. 2 Psorophytum 


_ \4 (Balearic Is. 
(vig ( ) 
/ 
j H.socotranum = 


vib 
vi ) / (Socotra) 
/ 


H. revolutum 
subsp. revolutum 


(Yemen to Cape Prov., 
Cameroons, F. Poo 


(iib) 
H. roeperanum 


(Ethiopia to Transvaal, 
Angola, W. Africa) 


i co lk 
U 
1 
! 


a 


(viic) 


H. quartinianum 


Yemen, Ethiopia to Zambia 
& N. Mozambique) 


( 


(vii) 
an dA 
(v) (ivy 
Sect.29 Brathys 4 
(Gr. Antilles, Belize, wu 7 
= 4 


Costa Rica to Bolivia) <a 


revolutum subsp. 


keniense 
(E. African mts.) 


H. 
(ii)-7 
Sect.30 Spachiuma« 


America, trop.& S.Africa, 
E.& S.E. Asia, Australasia) 


(s. 


Sect.22 Arthrophyllum 
(Levant, S. Turkey) 


7 


Sect.25 Adenotrias 


a 
oi (Morocco,Mediterranean) 


rd 


Sect.24 Heterophylla 
a. Turkey) 


Sect.23 Triadenioides 


(Socotra , Levant, S.Turkey) 


| / 
! FAN a) 
) 


'7 Sects 4-19 


' / (Europe,W. Asia,N.Africa) 


/ 
- H.madagascariense 
I (Madagascar) 


Sect.3 Ascyreia 
& E.Asia, W.Malesia, 
N. Turkey) 
H. lanceolatum 
subsp. lanceolatum 
(Réunion,Comoro Is.) 


4 
4 


2 H. lanceolatum 
Gii)_y subsp. angustifolium 
(Réunion) 
l... 
iaiia) 


Sect.26 Humifusoideum 


(New Guinea,Philippines,W. Malesia,trop. 
& S.E.Africa ,Madagascar) 


H. bequaertii 


S.steyermarkii 
(Mexico,Guatemala) 


| [Hypericum ] 


[ Santomasia | 


Fig. 61 
butions and their connections with other sections (cf. 


discussed in the text (but see footnote, opposite). 


Relationships diagram of the species in sect. 


(Ruwenzori Mts.) 


—— Sect. Campylosporus 


—--- Other Sections 


1. Campylosporus, showing their distri- 
Fig. 3). Roman numerals indicate trends 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 187 


In the other group related to H. quartinianum (sects 27. Adenosepalum and 
28. Elodes), however, the styles are separate. The two most primitive species of sect. 27 
(H. glandulosum and H. reflexum) are both natives of the Canary Islands and Madeira. 

H. roeperanum, also, has given rise to groups with partly or wholly free styles. In 
H. gnidiifolium, a localised Ethiopian endemic, they are free only near the apex; but in 
the monotypic sect. 21. Webbia (H. canariense), from the Canary Islands and Madeira, 
they are completely free and separate at the base, and the inner floral whorls are 
trimerous.* From it is derived the eastern Mediterranean group of dwarf shrubs 
comprising sect. 22. Arthrophyllum. 


(b) Sect. 3. Ascyreia and its derivatives. I. Sects 3-8 (Fig. 62) 


Sect. 3. Ascyreia has its most primitive species, Hypericum mysurense, in the extreme south 
of its range, in Sri Lanka and south India, and immediate derivatives in Orissa (H. gaitii) and 
Nepal (H. podocarpoides N. Robson, H. cordifolium) (Robson 1977b). From Nepal, the 
distribution of the section is effectively continuous westward to the Afghanistan border (H. 
oblongifolium) and eastwards to lowland China and northern Taiwan (H. monogynum). 
There are also two disjunct areas: (i) in north-western Turkey and adjacent Bulgaria is H. 
calycinum, which is most closely related to the western Himalayan H. oblongifolium, and in 
Indonesia, from Sumatra to Flores and south-western Celebes, are H. leschenaultii and (in 
north Sumatra only) H. uralum. The latter also occurs in the central to eastern Himalaya, 
whilst H. leschenaultii is most closely allied to an undescribed species from northern 
Thailand. 

Sect. 4. Takasagoya comprises four species, which are confined to Taiwan except for H. 
geminiflorum, which reaches northern Luzon. The most primitive species, H. formosanum, 
is most closely related to H. monogynum, but to the Szechwan form, not the one that occurs 
on Taiwan itself. 

Sect. 5. Androsaemum also comprises four species, of which the most primitive is H. 
grandifolium of the Canary Islands and Madeira. Its nearest relative in sect. Ascyreia is H. 
griffithii, which is endemic to Bhutan. The large gap between the Canary Islands and Bhutan 
is to some extent bridged by two of the other species, H. hircinum and H. androsaemum, 
which have dissected distributions from western Europe and adjacent Africa to northern 
Iran. H. hircinum, which in general is the more southern, has also been found in south- 
western Saudi Arabia (Asir). 

The monotypic sect. 6. Inodora (H. xylosteifolium) is confined to the high-rainfall area of 
Georgia and adjacent Turkey. Like sect. Androsaemum, its nearest relative is the central 
Himalayan H. griffithii. 

Sect. 7. Roscyna has four species, of which three occur only in central and western China. 
The most primitive, H. elatoides, is most closely related to H. monogynum and is confined to 
southern Shensi; H. pedunculatum occurs in Shensi and Hupeh; and H. przewalskii is spread 
more widely, from Shensi westward to Tsinghai and south to Yunnan. The fourth species, H. 
ascyron, is very variable, but it does not seem possible to recognise any subspecies or 
segregates such as H. gebleri Ledeb. It stands apart morphologically from the other three 
species and has much the widest distribution of the four. In eastern Asia it occurs from 
Yunnan and northern Vietnam northward to Manchuria and westward along the mountains 
skirting Mongolia to the Altai Mts and north-eastern Kazakhstan, including southern 


*At a late stage in the preparation of this paper, I realised that H. gnidiifolium was derived from H. roeperanum, not 
H. revolutum. Its markedly discolorous, densely reticulate-veined leaves and 3—5-flowered inflorescences are consis- 
tent with this hypothesis; and its incompletely united styles can be interpreted as an intermediate stage between the 
usually completely united styles of H. roeperanum and the free ones of H. canariense (sect. 21. Webbia). H. 
gnidiifolium thus belongs to trend vii c, not vi b as shown in Fig. 61. In making this alteration, which does not affect 
any of the phytogeographical argument, I am agreeing in part with Moggi & Pisacchi (1967); but, unlike these 
authors, I prefer to retain H. gnidiifolium as a species distinct from H. roeperanum. 


188 N. K. B. ROBSON 


/ ~ Ly F uy 
1 or a lbs, 24 5 Salle dee decile i is ae 
/ > ‘i =i 


\ 


EQUATOR 


Ja 
/ yh ME LZ 
X14 4-8-7 REDRAWN by M.GRIERSON after H.R.HEWER ~—»--1----4 
H S 


- ROYAL BOTANIC GARDENS. KEW 


‘ 
MOLLWEIDE'S PROJECTION 


C.B.H. 7510 - Dd, 208276 - - 7/67 


Fig. 62 Distribution of sect. 3. Ascyreia and four of its immediately derivative sections: 4. 
Takasagoya, 5. Androsaemum, 6. Inodora and 7. Roscyna.®) = centre of distribution of sect. 1. 
Campylosporus; x = areas of most primitive species of sects 3, 5 and 7. 


189 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 


= 60° 
/ 


ss 
\ N 
I N 
i N 
i 
| SES 
: 1 \3 
a +4 y\ 
b =< ean 
Al b ay ' 
i yh | \ 
: 1 roa 
f My ? Pore ! \ 
AS Re eal Me eee Sa ee ee aera Wyse---+-7 
/ | ‘ % 7 
| Ar 
y 
/ Si g 
I 
I 
\ 
+ 
' 
H 
H ot” 
| Fe | 
i | 
1 | 
| ae 
I *2- oats 
| Leen y 
| 
=49) | 
\, 
iy t 
hha? Wt ee N 
Lay 
| os 
i : 
\ 
Dine 
Sot i 
are i} / 
ie 
AG 
wh 
ae ° a 
Pe o | a 
We wit 
| ae i a a 
7 : 
| i ! : 
Z P | 
* ‘Dae eae F 
fe ae: o 
7 -~ | 
A pad | 
us | Pig | 
r4 a | 
4-2 | 7 | eo 
4 - Be yl | h Ci dell z ee 


N. K. B. ROBSON 


190 


Ss5 
=< 
OL 


x 


--~«-- 


Zar 


| 
| 
\ \ 
: Sk 
\ as 
\ | 1 
\ | SS a | 
‘\ é St | 
. 7 
\. 
N eo yl! 
Sel —— 
~ 
Sal ape aiaa sr Senay = 
ae | 8. 5 
aN | 
N | 
SS | 
<< | 
ES | 
> 
~ | 
AL 
Sus 
a 


pt 


~ 
~ 
> 


~. 


ROYAL BOTANIC GARDENS. KEW 


- 
- 
—_- 


Lf 
Se ll 


~~ 


pyeree S ns H 
S i 1 
1 1 
1 4 
al 
ier J 
ee vy 
ee al fe 
ec 1 end 
-—- i Ze 7 
| yf 
| 7 
Weal e 
ee 
“7 Il ee 
oa [ee 
Le ae 
oe 7 
ese 
ca 
oe i-4 
Za Ww 
ig = 
a w 
= 
se 
x 
. 
o 
KM 
< 
oO 
Zz 
[e} 
a) 
[- 4 
SA Ww 
Pee a 
~ oO 
~ d 
ENG = 
~ 
— mS 
“Sy aN 7 
~ ~ 
~ teas 
1 N cS 
St \ = 
K & < 
eas. a 
INN. 
| Ne ON, 
| By N 
| PNG 
| Oras 
+ SK 
| NA 
| Y 
1 1 
1 wll 
\ sy 
I >| 
I 
ete | 
| af ae 
\ i 
| 4 
ee eaeee el 
peer 
| S oe V/ 
| Bea y 
! ea Ss 
Pec fe 
i. 
VL yas 
SS 1 eo 5 
ee | ie 4 nH 
| | ly 3 
| 6, = 
H | Did 
- “ww 
| ae Ww 
| ai 3 
| Z~ ao 
i} PL ere) = 
I Sa +r = 
= a 
Cea ce) 
Ss = 


- 7/67 


63 Distribution of sect. 9. Hypericum. 


® (Mexico) 


C.B.H. 7510 - Dd, 208276 - 


area of H. graveolens. 


area of H. pedunculatum (sect. 7); 


(China) 


® 


centre of area of H. formosum; @&(eastern U.S.A.) 


Fig. 


191 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 


/ 


Se ===] 40° 
/ 
/ 
Jo0° 


| 
| 
tl 
Ix 
A 
AG 


- 


See 
oy ae 


oe 
| 


! 
/ 
~---40}------- 
/ 
7780 
7 


+ 

i} 

1 

| 

| 

\ 

| 

| 

! 

i} 

| 
ae 
& 

, 

if 

| 

1 cand 
~! 


: I 
I 6 4 ! 
J I 4 Het 
O bite peat 
¢ ie 37 
ro i = We 
eee | eo 
| 2 
oe 
ae 
- 
isos 
| a 
| | Sa 
@ 4 East 
eae OE Ee ee a el fe a he eles ee ~ 
(ure Poe | te aS 
2? 1 | | Sree ' Se 
+o TS) eS 
4 lo \ ~~ BS 
Ay | | tie eS 
Hone ; | | : | en 
| a | % 4 | IN 
ete ee a See ce ae Cor as St Ss we =a = = = = at 2 ale 
H | 1 | a ee ul SN 
| | i | } ane as 
| ey : | eT eS 
N H | | | : =| 
1 | ! = wl ! 
| I | | eel | 
| | : | ee ae 3 # 
a z —————— on ee | | 
Litege® 5 Sees soa lee ay 
Ve | ! i Bey 
| \ | | ee vA 
l | _ =| 1 Lat oA 
\ SG £ ol | ee we 
| od I tb ee L- 
= 09} < Sf |e Spates Bee 
ee | = = ae Boe = - 1 Ler 
Tod--------2s8=3 SSS ee H eo 
ve=-- | see 
gor | all 2— 
27 | 228 
oo fis | -<===T__ 
a | : aa ee | 
Be | | o f ae 
Fal ‘ | a i 
an So I 1 eS 
oo | | ed) ——e ! Ss 
Pie pe g H | 
/ i cae oe 
ve ! she | ! 
Z | Pe I 
/ Liege | | | 
76, Ve | | 
re wa = | a? | 


192 N. K. B. ROBSON 


Kamchatka, Sakhalin, Japan and Taiwan (one record). It also occurs in the north-eastern 
part of the U.S.A. and adjacent Canada. The American form, with pyramidal branching (H. 
pyramidatum Aiton), can be matched in different parts of the Asian range (e.g. Yunnan, 
Japan). 

As well as being the origin of the large sect. 9. Hypericum (see p. 190, Fig. 63), sect. 7. 
Roscyna has apparently given rise to two monotypic sections. Of these, sect. 8. Bupleuroides 
(H. bupleuroides) is related to H. pedunculatum and H. przewalskii but is separated from 
them by the central Asian plateau and steppe region. It occupies almost the same area of 
western Transcaucasia as does sect. 6. Inodora. 

The other species, H. concinnum, which was included in sect. Hypericum in Part I 
(Robson, 1977a), is the only member of Sect. 9a. Concinna (see p. 173). It is confined to the 
coastal range and Sierra Nevada of California from the San Francisco region northwards, and 
thus, like American H. ascyron itself, is separated from its relatives by the north Pacific 
Ocean. 


(c) Sect. 3. Ascyreia and its derivatives. IT. Sect. 9 (Fig. 63) 


The major derivative of sect. 7. Roscyna is sect. 9. Hypericum, a large section with a wide 
north-temperate distribution. Like sect. Roscyna it has its most primitive species in eastern 
Asia or, at least, centred there, and all members could be derived from a plant resembling H. 
pedunculatum (sect. Roscyna, from Hupeh). It is not possible to nominate one species as the 
most primitive. H. yamamotoi (from Hokkaido, Japan) most nearly resembles an ancestral 
species in some respects. It is related to (i) an east Asian group (Kamchatka to Luzon, 
northern Vietnam and central Nepal, with outliers in Sumatra and Borneo) and (11) a mainly 
west Asian and European group (H. maculatum, from Siberia to western Europe, and its 
derivatives; H. attenuatum, from north China and adjacent Siberia, and H. elegans, from 
central Siberia to central Europe; H. perforatum, which looks and behaves like the result of 
an allopolyploid cross between H. maculatum and H. attenuatum). (ii1) In addition, there is a 
North American group, which comprises two subgroups: H. formosum, in northern and 
central Mexico, and its derivatives in Mexico, the U.S.A. and adjacent Canada (H. scouleri, 
H. punctatum, H. pseudomaculatum) and the H. oaxacanum affinity, mainly in central and 
southern Mexico and adjacent Guatemala. The most primitive species in the latter subgroup, 
which occurs from southern Vera Cruz to Guatemala and is close to H. oaxacanum, is as yet 
unnamed. It has another very close relative, however, in North Carolina, H. graveolens, 
which appears to have hybridised with a species from the other subgroup, H. punctatum (see 
p. 167). Although neither H. formosum nor H. sp. aff. oaxacanum can be regarded as wholly 
primitive in the North American group, a plant possessing the primitive characters of both 
species would be very near H. pedunculatum and H. yamamotoi from eastern Asia. 


(d) Sect. 3. Ascyreia and its derivatives. III. Sects 10-16 (Fig. 64) 


The most primitive section of the Olympia group (sects 10-16), from which all the others 
have apparently been derived, is sect. 10. Olympia, which has two disjunct areas: 1) 
Hypericum olympicum p.p. in Greece, adjacent Yugoslavia and Bulgaria, north-western 
Turkey and some northern Aegean islands; ii) H. olympicum p.p. and H. polyphyllum in 
southern Turkey from Antalya to the Amanus Mountains (Robson, 19804). The most primi- 
tive member of the section, H. polyphyllum subsp. lycium Robson & Hub.-Mor., which is 
confined to a small area near Takhtali Dag in south-western Antalya, has its nearest 
ancestral relatives in Nepal (H. cordifolium and H. podocarpoides N. Robson, sect. 3. 
Ascyreia). 

Sect. Olympia has given rise to two groups, one eastern (sect. 12. Origanifolia in Turkey 
and western Transcaucasiz, with the derivative sect. 11. Campylopus extending north- 
westwards into Greece and Bulgaria) and the other primarily north-western (sect. 13. 
Drosocarpium). Sect. Drosocarpium has its most primitive species (H. rumeliacum) centred 


ft 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 193 


Fig. 64 Distribution of Olympia group (sects 10-16);—= sect. 10. Olympia, «e+. sects 11, 12, 15, 
16;==sect. 13. Drosocarpium; =-=sect. 14. Oligostema. x = area of most primitive taxon of sect. 
10—AH. polyphyllum subsp. lycium. 


in the southern Balkan peninsula (Makhedonia, Thrace), and extends eastward into Trans- 
caucasia (H. bithynicum, H. montbretii), southward into Crete (H. trichocaulon, H. kelleri) 
and Cyprus (H. repens*), westward throughout the Mediterranean and north-west Africa to 
Madeira and the Canary Islands (H. perfoliatum) and northward to the Balkan mountains, 
Carpathians, Alps and Pyrenees (H. richeri sens. lat.). 

| Sect. Drosocarpium, in turn, has given rise to three sections: Sect. 14. Oligostema (in the 
western Mediterranean, Madeira and the Azores, with H. /inarifolium also in Atlantic 
Europe and H. humifusum extending to northern Scotland, southern Scandinavia, western 
U.S.S.R. and the northern Balkans), sect. 15. Thasia (H. thasium in Thasos and the adjacent 
mainland) and sect. 16. Crossophyllum (H. adenotrichum in western Anatolia, H. orientale 
in northern Anatolia and adjacent Transcaucasia). 


(e) Sect. 3. Ascyreia and its derivatives. IV. Sects 17-19 (Figs 65, 66) 


This second western Eurasian group (the ‘Hirtella’ group) seems to have been derived from 
the same part of sect. 3. Ascyreia as has the Olympia group, viz. the Nepalese Hypericum 
cordifolium and H. podocarpoides. However, whereas the Olympia group (with basic 
chromosome numbers n=9-7) is centred in western Anatolia and has primitive species 


*H. aucheri and H. repens were included in sect. 14. Oligostema in Part I (Robson, 1977; also in Robson, 1968) on 
account of their vittate capsules; but their other characters all indicate that their affinities are clearly with sect. 13. 
Drosocarpium (with H. rumeliaacum and H. trichocaulon, respectively), to which they must therefore be transferred. 


60 


194 N. K. B. ROBSON 
: | <A 
0°’: ch A) 
poe Or 
| 
o 
a\ f° 
og 
A 
ey 
] 
co 4 = 
° a eo : ° 
ay Ena Te = - : 
a Se i 


Fig.65 Distribution of sect. 17. Hirtella. x = area of most primitive species. 


which are morphologically closer to H. podocarpoides, the Hirtella group (with basic 
chromosome numbers probably n= 10-8) is centred in eastern Anatolia and Transcaucasia 
and has primitive species which are morphologically closer to H. cordifolium. 

From eastern Anatolia, species of sect. 17. Hirtella (Fig. 65) extend round the Fertile 
Crescent south-westward to Israel and Jordan and south-eastward along the Zagros 
Mountains to Shiraz in Iran. Of the four species that occur fariher east than Khrebet Dag 
(U.S.S.R.-Turkmenia), two (H. elongatum and H. scabrum) reach the Altai Mts (U.S.S.R.- 
Kazakhstan). Beyond the Caucasus-Anatolian region, where by far the greatest number of 
species are found, the section ‘peters out’ westwards with two markedly disjunct species 
which have distributional ‘arcs’ to the south and north, respectively, of Italy. To the south, 
the mainly Central Asian to Anatolian H. elongatum occurs in central Greece (as H. 
tymphresteum Boiss. & Spruner) as well as in Morocco and south-eastern Spain (as H. 
callithyrsum Cosson). This species has also been recorded from the Crimea, as has H. 
hyssopifolium; but all the specimens from there that I have seen belong to H. lydium (€ H. 
ponticum Lipsky). To the north, H. hyssopifolium sens. str., which approaches the primitive 
species of sect. 19. Coridium (H. asperulifolium), is found in the central Balkans (Serbia, 
Bulgaria), the south-western Alps (France, Italy) and eastern and south-eastern Spain, where 
it meets H. elongatum. 

Sect. 19. Coridium (Fig. 66) comprises five species, four of which are morphologically and 
geographically isolated. Of these, three form a morphological and geographical series: H. 
asperulifolium (eastern Transcaucasia), H. coris (central and western Alps) and H. ericoides 
(south-eastern Spain, Morocco, Tunisia). The other two are closely inter-related but more 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 195 


eae 


Fig. 66 Distribution of sects 18. Taeniocarpium —— and 19. Coridum === x=area of most 
primitive species in sect. 18,)= area of most primitive species in sect. 19 (H. asperulifolium). 


distantly related to the first three: H. empetrifolium (N. Albania, Aegean region, Cyrenaica) 
and H. amblycalyx (Crete). The recently described Cretan H. jovis Greuter is very similar to 
the latter and may be conspecific with it. 
| The other section in the Hirtella group, sect. 18. Taeniocarpium (Fig. 66), has the most 
_ primitive form of its most primitive species, H. linarioides (i.e. that named H. polygoni- 
_ folium Rupr.), in Transcaucasia. Thence the section extends round the Fertile Crescent 
south-westward to Israel (Mt Carmel) and eastward to the Elburz Mts. It is present through- 
out Anatolia except the extreme west; but, apart from two widespread species, it is otherwise 
confined to isolated areas round the north of the Mediterranean and Black Sea: in Greece (H. 
taygeteum—Taiyetos Mts; H. fragile—Attica, Evvoia), the central Balkans and the Crimea 
(H. linarioides), Albania (H. haplophylloides), and the Pyrenees and western Alps (H. 
nummularium). One of the two widespread species, H. hirsutum, has a mainly ‘continental’ 
distribution, from northern Iran westward to Algeria, Spain and eastern Ireland, and north- 
ward to beyond the Arctic Circle in Norway and eastward to central Siberia (R. Yenisei 
valley). The other, H. pulchrum, has an ‘Atlantic’, north-west European distribution, from 


northern Portugal to eastern Germany, southern Norway and the Far6ées, with scattered 
localities further east. 


(f) Sect. 20. Myriandra (Fig. 67) 


The most primitive species in sect. 20. Myriandra, Hypericum frondosum, is restricted to a 
relatively small area in south-eastern U.S.A. around the Cumberland Mountains (east and 


N. K. B. ROBSON 


196 


4 
fe} 
= 
=< 
=) 
fo] 
Ww 


Sho ee 


M.GRIERSON after H.R.HEWER 


ROYAL BOTANIC GARDENS . KEW 


- 


REDRAWN by 


= 


i 


eee ee 


- 10/64 
Fig.67 Distribution of sects 20. Myriandra Zand 27. Adenosepalum:*** H. glandulosum group, 


C.B.H, 4623 - Wt. 34154 - Dd. 10 - 


areas of most primitive species. 


& H. reflexum group. x 


oy 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 


4+-------- 


a 
—- 
~ 

=: 
\ 
N 
XN 


2 a Dl oa 
a? 3 ea 


n 


180° 


1 
ip 
Ip 
| 
/ 


—740° 
/ 


/ 
--————-f-—— 
/ 
/ 


PROJECTION 


MOLLWEIDE S 


Nee 
1 


~ ” 
~— _— 
---L=— 


198 N. K. B. ROBSON 


central Tennessee, central and south-western Georgia and northern Alabama). In its meristic 
variation it foreshadows two of the major trends which the section displays, viz. pentamerous 
(‘Hypericum proper’) and tetramerous (‘Ascyrum’) perianth whorls (Figs 4, 5). The 
‘Ascyrum’ group 1s more south-eastern in distribution, with only H. crux-andreae (H. stans)* 
occurring as far north as New Jersey; the area of one of the mainly Florida species (H. 
tetrapetalum) extends into Cuba, whilst only H. hypericoides subsp. hypericoides occurs 
south and east of Texas (Mexico to Honduras Republic) as well as in Bermuda, the Bahamas 
and the Greater Antilles. Of the essentially pentamerous species, the ‘Myrtifolium’ group 
(which includes the tetramerous H. microsepalum—Fig. 5) is wholly North American, with 
the area of H. ellipticum extending north to Lake Superior and Newfoundland. The other 
pentamerous ‘branch’ in Fig. 5 comprises the central to northern ‘Prolificum’ group, which 
has given rise to a more southern, narrow-leaved series, the ‘Galioides’ group. The latter is 
centred in the south-eastern U.S.A., with H. nitidum occurring also in Cuba and Belize (but 
not Mexico) and giving rise in Cuba to H. limosum Griseb. 


(g) The Macaronesian and north-east African links with the Mediterranean (Sects 2, 21-25) 
(Figs 68-70) 


Sect. 21. Webbia, in my view, comprises one variable species, Hypericum canariense, which 
occurs both in the Canary Islands (all except Fuerteventura and Lanzarote) and in Madeira 
(Fig. 69). Its nearest ancestral form is to be found among the representatives of H. 
roeperanum and H. gnidiifolium in Ethiopia, and it, in turn, appears to be ancestral to a 
group of five species in the Levant and southern Turkey (sect. 22. Arthrophyllum) (Fig. 69). 
H. balearicum (sect. 2. Psorophytum) seems to be related directly to the north-east African 
complex of H. revolutum (Fig. 70). 

The island of Socotra is rich in endemic Hypericum species, which form a morphological 
link between the east tropical African H. revolutum subsp. keniense (sect. 1. Campylo- 
sporus) and several Mediterranean species. The H. socotranum group itself is very variable 
and consists of three distinct entities (species or subspecies) in Socotra. Also on this island 
and related to one member of the H. socotranum group are the primitive members of 
sect. 23. Triadenioides, three species in all, which are clearly closely allied to two species 
from southern Turkey and the Levant, H. ternatum and H. pallens (Fig. 68). 

H. socotranum sensu stricto is the nearest relative of two other mainly Mediterranean 
sections, sects 24. Heterophylla and 25. Adenotrias. The monotypic sect. Heterophylla (H. 
heterophyllum) is confined to a small region of western Anatolia, whereas the members of 
sect. Adenotrias occur in scattered areas over the Mediterranean and north Africa, mostly on 
islands (Figs 68, 69). The most primitive forms of H. aegypticum are found in southern 
Morocco, and the forms become progressively more reduced in an eastward direction. The 
other two species are in the east, H. russeggeri in Syria and southern Turkey, H. aciferum in 
Crete (Figs 68, 69). 

In view of this prevailing eastward trend, it seems possible (though unlikely) that H. 
heterophyllum has been derived from H. aegypticum, i.e. that it has reverted to homostyly 
rather than that their ancestral lines diverged before the evolution of heterostyly. 


(h) Old- World transcontinental distributions I. Sect. 26. Humifusoideum (Figs 7, 70) 


The distribution of sect.26. Humifusoideum provides a considerable interpretation 
problem. The most primitive species, Hypericum sewense, has been collected only once, 
from Madang District in north-eastern New Guinea; but the other four species of this section 
in New Guinea are closely related to it. The group most closely related to these species, 
however, comprises the Madagascar-Mascarene members of sect. 1. Campylosporus, of 
which H. lanceolatum subsp. angustifolium is most similar morphologically to H. sewense 
(Figs 3, 60, 61). 


*For an explanation of this synonymy, see Robson (1980c). 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 199 
30° 50° 


ssa x 


’ 


at Stay) \ 


Fig. 68 Distribution of sects 23. Triadenioides == 24. Heterophylla ——_and 25. Adenotrias 
(excluding H. aegypticum in part) eceee . 


"JOUITXO = X ‘Ma SOPO]T “QZ pure (A[UO °%¢ wintOf14dDI "YY pue *="°= WiN]JDUNPVOD ‘F7) Wn]Ddasouapy “L7T* *** (Satseds 
12) Spiouapy ST * == DIEAYdosajay] ‘pT ‘= == SaplolUapviay, “CT “mmm UNA Y AOAYIAP “ZZ “meee =DIGGAM “17 $1028 JO UONNGINSIG 69 “BIH 


N. K. B. ROBSON 


200 


————————— 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 201 


The derivative species of the New Guinea ‘core’ of the section are in three widely scattered 
groups: (i) H. pulogense in northern Luzon (Philippine Is.), (ii) H. beccarii, with two sub- 
species respectively in central Sumatra/west Java and northern Sumatra, and (iii) three 
species in Africa—H. natalense in eastern Cape Province, Natal and Transvaal, H. wilmsii in 
the same general regions and also eastern Zimbabwe and central Madagascar, and H. peplidi- 
folium from eastern Zimbabwe north to Ethiopia and west to Angola, the Cameroon 
Highlands and Fernando Poo (Robson, 1958), 1961, 1973a, 1974; Bamps, 1971 Killick & 
Robson, 1976). All three groups seem to be related directly to the New Guinea one, H. 
natalense being particularly similar to the widespread New Guinean H. papuanum. 


(i) Old- World transcontinental distributions. II. Sects 27-28 (Figs 67, 69) 


Sect. 27. Adenosepalum comprises two groups (subsections ?), both of which have their most 
primitive species in Macaronesia. The species nearest to sect. 1. Campylosporus is 
Hypericum glandulosum (western Canary Is.; Madeira), to which is allied a group of tropical 
and southern African species which are completely glabrous (Robson, 19585, 1980a). Like 
the whole section, which has a markedly dissected distribution, these African species occupy 
widely scattered areas: H. kiboense in east Africa, H. conjungens in the Lake Tanganyika/ 
Lake Nyasa area, H. abilianum in Angola (Huila) and H. aethiopicum (two subspecies) in 
the Zimbabwe/Mozambique border area, Angola (Huila) and South Africa (Robson, 1980a). 

The other group stems from H. reflexum sens. lat. (western Canary Is.), a shrub with 
pubescent stems, and its relatives are probably primarily pubescent, 1.e. there is evidence 
that the wholly glabrous species are secondarily so. They derive more immediately from H. 
annulatum, a variable (rarely glabrous) species with its most primitive form in Ethiopia/ 
Eritrea (“H. intermedium’) and apparently conspecific populations in east Africa (‘H. 
afromontanum’), the Balkan Peninsula (‘H. degenii’) and Sardinia (H. annulatum sens. str.). 
H. annulatum is very closely related to H. montanum (central and west Europe, north 
Africa, the Crimea and western Transcaucasia) and some eastern Mediterranean species. 
Other Mediterranean, north African and Canary Island species show direct connections with 
the north-west African and western Mediterranean H. pubescens or its Somalian relative H. 
somaliense, the trans-Saharan gap being bridged to some extent by H. psilophytum, which 
occurs in the Hoggar Mountains (south Algeria). 

The Asian species of sect. Adenosepalum, also, are related directly to H. annulatum, but 
they are wholly glabrous. The most primitive species, H. elodeoides, is distributed along the 
Himalayan range from Kashmir to western Assam, with disjunct populations in Khasia, 
Yunnan and Assam/Burma; and derivative species occur outside that area in China 
(Kweichow), northern Thailand, Tamil Nadu (Madras) and Sri Lanka (Robson, 19775). 

As I have explained earlier (Robson, 1972a : 378), H. elodes (sect. 28. Elodes) is 
morphologically very close to sect. 27. Adenosepalum, if the floral specialisations are 
ignored. Its nearest relative in sect. Adenosepalum is clearly H. coadunatum, which occurs in 
the Canary Islands (Gran Canaria) and also in Morocco and Algeria. The Gran Canaria 
plants (H. coadunatum sens. str.) are the most similar to H. elodes, which has an ‘Atlantic’ 
distribution in western Europe and the Azores (Fig. 69). The north African H. coadunatum 
(= H. naudinianum Cosson) on the other hand, is scarcely distinct from H. caprifolium of 
south-eastern Spain. 


(j) Sect. 29. Brathys (Fig. 70) 


The most primitive species of sect. 29. Brathys, Hypericum styphelioides, occurs in southern 
Belize and Cuba, the form in Belize being less specialised. The most closely related deriva- 
tive species (H. phellos, H. magniflorum and one hitherto undescribed) are all found in the 
Colombia-Venezuela border region, whence the section has spread along the Andes in both 
directions, east to Northern Venezuela (H. caracasanum) and Mt Roraima (H. roraimense) 
and south-west to Peru and Bolivia (H. andinum). Three species otherwise almost confined 


N. K. B. ROBSON 


202 


\ ane a XY 1 a GOS ee 1 1 
XN ae 1 
ES RE tree | person 
> ae — 
SS, = | =) | 
x \ See a 4-- a a) | 
»S t¢ T= \--oL ee ee ee | 
N Sl { | 
a ae | | | en . 
\ SS I ie! H Ae X | 
\ i XN 8 1 | | ae 
\ ae) I! | | we a 
\ I39 > \ pe a 
; -- = 
eee 5 
eae ey 
7 ie = 
| a, =3 A 
~ 
| ees Z 
1 | Tiss 3 
| | ° 
ie pee ! ‘| 
ia s 
5 


S 
= 
v 
vd 
S 
4 
= 
vat » : >: : 
i f ss 25 iy 2 a 
| eo : lon 
/ Re hee ee a ee a | | a <Q 
A / ee aa H ! 7 io) 
ip | ee « ae 
Wes | bs 
Kv | we $ ea 
ft ‘ | | Bee ‘fi SS No) 
io le { | ee ae SN 
ist, ' Jee oa ea 
it | re me ae =o 
3 | (oe . Si, oO 
t mK of) Se ee o S n 
' 4a) A 4s ie ee Sa ee a ee, a Qe SS Pas a LE 
| tle ee 3 ou 
f o 22 eee 5 ~G oO 
- 22 nS) 
r-4o 5 2e 
eae 28 pee 
i a 2.2 
iss a ee 
\ = 
\h 72 £ 
\ 35 Bue 
oN \ Z2 2 O82 
3 \ e, s Se} 
PON as : es fate 
. ae 1 2 
. gf 
cre 
nee 
Q n 
\ a 
| re 
San * Oa 
1 \ all 
2 i ee ee es 
cia. = 
a = 
pti | J 


203 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 


\ 
\ 
“9--\20° 


=. 
------+ 


“0 


180° 


PROJECTION 


—7/60° 
, 
S 


aa Sa 
/ 
7 
4 
MOLLWEIDE 


N. K. B. ROBSON 


204 


— 
~ 


~ 
= Ses 
~ 


Ao™. 


| 
pe & 
ia i 
lee = 
ry a 
ae r 
i & 5 
a me 
1 -~ x 
! ! Zs 
ae TS 2 3 
| of oo Ss, wis 
\ Oo i} mS & 
~ | . Ox 
4 “ 7 Zz 
\ K tr \ =< 
\ I| > a N S 
\ x ! ry 
\ N SS 1 ® 
he I g 2 
| | 
VAS é 
\ \\ SN ! w 
Nin N + m 
| 
\N \ | 
\ \ \\ 
a \ ) H \ 
ay \ | V\ 
NWO | 1\ 
\ \ a Si 
\ 
nM 
\ \ I 1 1 
\ | 1 
\ | | Sl 
\ f | ioe 
\ d} | | en 
i \ food 
\ | | i 
\ \ ily A i 
‘ \ i ty 
\ / 
\ \\ 1 y 
\ \ ! oS 
\ \ 73 
1 
SUI en ! 
\l SK | z 
oN S| £ 
\ ~~ 
\ | +- 9 
\ | w 
\ SK fo) 
\ » m4 
\ “ | a 
\ N @ ° | 
\ | XN ° ! a) 
\ | “ ° | 7 
XN é) NS ar ] a 
ee): : 
Se igs 3 
\ f 3 
N! | 8 
a | 
>. | 
“< 
PSS | 
sa 
coe 


- 7/67 


Distribution of sect. 30. Spachium. x 


C.B.H. 7510 - Dd. 208276 - 


® =apparently 


area of most primitive species; 


Fig. 71 


recent introductions. 


205 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 


8 
~ 
ip 
1 N 
1 ~ 
1 ~S 
I 
WY 
ten SN 
~ \ 
eta 
! al 
I : ie 
bi o H ! 
-----2.--- Ayst-t- 
i ‘ 
Aaa 
I Gi Tay 
I We 
1 =_ 
1 a 
a aaa 
! 7 
a 
' “ 
eo 
Ras 
“‘ 
rh 6 TSS 
EL oes 
iets “— 
1 ~ TASS 
| Sse ~ 
I aS 
| sat 
ji Seaiie i 
Seay 
\ See 
ia) 
| Bee 9) 
aetna ¥ 
| 1 
1 An 
| “*} 
H aay, 
! in eA 
I me 7 
5 7 
aah of, 
ee 


pie See one Ihe 


206 N. K. B. ROBSON 


to Colombia and adjacent Venezuela, or their close relatives, occur in the mountains of 
Costa Rica; and three other species, also with Colombian affinities, are endemic to 
Hispaniola. 


(k) Sect. 30. Spachium (Figs 71, 72) 


The widespread, intercontinental distribution of sect. 30. Spachium does not appear to make 
sense until morphological trends are considered; and, even then, no one theory seems able to 
account for it. 

The most primitive species, Hypericum rigidum, occurs in southern Brazil and has its 
most primitive form in Parana. It is clearly derived from the African H. revolutum subsp. 
keniense or from near it, and in turn it has given rise to a closely related group of species (H. 
brasiliense, H. connatum, H. myrianthum, etc.) in the region to the south of Amazonia, from 
the coastal range of Brazil to the Andean foothills of Bolivia. A group of North American 
species, H. denticulatum and its allies, in the south-eastern United States, is related directly 
to Brazilian members of this affinity. 

Another derivative species of H. rigidum, H. chamaemyrtus, from the Andes of north- 
eastern Colombia, provides the key to the sectional relationships, as several trends appear to 
diverge from it. Of these derivative groups, only one species pair, H. piriai and H. 
hilaireanum, occurs in the above-mentioned southern area (in south Brazil and Uruguay). 
Although somewhat isolated morphologically as well, it is related to H. gnidioides and its 
allies, which extend in area from the Honduras Republic, where the primitive forms are 
found, south to Panama and north to Mexico and the Revillagigedo Islands. H. gnidioides, in 
turn, has derivative groups in the Greater Antilles (H. dichotomum, H. fuertesii, H. 
diosmoides) and the eastern United States (H. drummondii, H. gentianoides). 

H. silenoides, a very variable species of the Andes from Venezuela to northern Argentina 
and the Galapagos Islands, is also directly related to H. chamaemyrtus through its most 
primitive forms, in southern Ecuador and northern Peru. It has given rise to H. caespitosum 
(H. brevistylum Choisy) of central Chile, which, according to Rodriguez-Jiménez (1973), is 
adventive in its north Chilean locality (Arica). The other Chilean coastal representative (H. 
paposum I. M. Johnston), however, cannot be differentiated from the protean H. silenoides. 
It is also very close indeed to the primitive forms of the E. Asian/Australasian H. 
gramineum, which are found in New Zealand, Tasmania and south-eastern Australia. This 
species shows three fairly distinct morphological and geographical trends (Robson, 
1973a: fig. 3): 

(1) New Zealand, New Caledonia, New Guinea, Taiwan. 

(2) New Zealand, eastern Australia, eastern Himalayas. 

(3) New Zealand, eastern Australia, western Australia, Vietnam. 

The last group derived from H. chamaemyrtus comprises two subgroups, one mainly 
American and one African. H. pauciflorum of central and northern Mexico is the basic 
species of the first subgroup, and the areas of it and its immediate derivatives (H. moranense 
=H. pratense Cham. & Schlecht. sens. lat.) extend north to Texas and south to Panama. 
The remaining species are derived from H. majus of the northern United States and Canada, 
viz. H. anagalloides (western N. America), H. canadense (eastern N. America) and H. 
gymnanthum (south-eastern United States with a disjunct Guatemalan population). From 
the last-named there appear to have arisen two widely distributed species, H. japonicum and 
H. mutilum. The former has its primitive forms in the Philippines, Taiwan and adjacent 
China and extends north to Japan (southern Hokkaido), west to the Punjab and Sri Lanka 
and south to New Guinea, extreme south-eastern Australia and New Zealand (Robson, 
1973a, fig. 5). The New Zealand form also occurs in Hawaii. The latter species (H. mutilum) 
is polymorphic and, although centred in the eastern United States and Canada, has disjunct 
populations in various parts of Central and South America and in the Azores, north-western 
France, northern Italy, Poland, Transcaucasia and Hawaii. In turn, it appears to have given 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 207 
H. rigidum gr. 
(S. Brazil to Bolivia, Argentina) 


nes H.denticulatum gr. 


(S.E. United States) 


H. piriai He pirial geet +H. H. chamaemyrtus ——pH. silenoides gr. H. silenoides ontlyH. gramineum 


G Brazil, Uruguay) (Colombia) (Venezuela to Argentina, (New Zealand, 
Chile, Galapagos) Australia [S.& E. Asia]) 
H.lalandii gf 
am (S.Africa to Angola, Sudan & 


H. gnidioides gr. H. pauciflorum gr. Nigeria) [Madagascar] 
Honduras Rep. to onduras Rep. to Panama (Mexico to Texas & Pana to Texas & Panama 
& a 


H. humbertii gr. 


(Rwanda to Ethiopia & SW. 


Fale Bhd S gr. Africa) 


(Gr. Antilles) H. gentianoides H. gentianoides gr H. perrieri gr. 
(NE America [Hispaniola, a as (Madagascar) 


Paraguay, W.Europe 


H. majus 
(N. United States, 


eon [Japan,W Eu dm 


H. anagalloides H. canadense 
(w. Canada, United States (E. United States & Canada, 
& Mexico) [w. Europe] ) 


H. gymnanthum 
(S.E. United States, 


Guatemala [c. Eu rope]) Sy 


H, japonicum H. mutilum 
(Japan to Srilanka & (E. & C.United States & 
New Guinea, [S.E. Canada {rest of America, 
Australia, New Zealand Europe, Caucasus, Galapagos, 
& Hawa) Hawaii, Azores] ) 


Intercontinental disjunctions 


Intracontinental disjunctions 


[ ] Intraspecific disjunctions 


Fig. 72 Evolutionary and geographical relationships in sect. 30. Spachium. 


rise to H. boreale (north-eastern United States and Canada) (cf. Gillett, 1979), H. arena- 
rioides (Cuba) and H. pleiostylum Rodr.-Jim. (eastern Brazil). 

All the species of this subgroup except H. anagalloides and the H. mutilum derivatives 
have also been found in isolated localities in western or central Europe, as has H. genti- 
anoides (cf. Heine, 1962; Robson, 1968)). 

The second subgroup related to H. chamaemyrtus is African. H. lalandii, which is distri- 
buted from Cape Province to the southern Sudan and northern Nigeria (Bauchi) and also 


208 N. K. B. ROBSON 


occurs in Madagascar, has its most primitive forms in Natal and the Transvaal. These are 
very similar to primitive Mexican forms of H. pauciflorum and a related species in the 
Colombian Andes, so much so that these three taxa are difficult to distinguish from each 
other. Within Africa and Madagascar, however, the variation in H. /alandii is great; and it 
appears to have given rise to distinct species groups in Madagascar (H. perrieri group) and 
east Africa (H. humbertii group). The latter extends north to Ethiopia and south-west to the 
Caprivi Strip. 


Interpretation of Hypericum distribution 
(a) Introduction 


From the geographical relationships depicted in Figs 60 and 61 it is clear that Hypericum 
must have spread from Africa to the adjacent continents or, rather, to the continents that 
were adjacent in Gondwanaland before that supercontinent began to break up. Hypericum 
and Santomasia can thus be interpreted as Hennigian sister-groups, one in the east and the 
other in the west of west Gondwanaland. The present relict distribution of Santomasia on 
either side of the Mexico-Guatemala border (in an area that was not part of Gondwanaland) 
could be regarded as secondary, i.e. to have been attained after North America made contact 
with South America. Alternatively, but less likely, it could have been attained via north 
Africa and eastern North America. 

At the outset of this analysis of Hypericum distribution, it is necessary to clarify two 
points, relating respectively to dispersal and time of origin: 

(1) The natural distributions of all sections of Hypericum can be harmonised with current 
theories of plate tectonics, with the apparent exception of sect. 30. Spachium. In general, 
therefore, it seems unnecessary to involve long-distance dispersal in the interpretations. 
Even though the seeds of Hypericum are small, there is no direct evidence of inter- 
continental transport by birds (cf. p. 124), wind or air-currents; the evidence of such trans- 
port in sect. Spachium, although apparently strong, is circumstantial. 

(2) There is no unequivocal fossil evidence for the existence of angiosperms before the 
Early Cretaceous (Hickey & Doyle, 1977; Hughes, 1977); and the earliest fossils that do 
exist are of monocolpate pollen grains, whereas those of the Dilleniidae (including 
Hypericum) are basically tricolpate. Some workers therefore deny that the angiosperms 
could have originated much earlier than the Cretaceous without detection; others, con- 
sidering their undoubted diversification into modern families by the Middle Cretaceous at 
least, believe that they must have had a long pre-Cretaceous existence (cf. Lakhanpal, 1976). 
To some believers of the former theory (e.g. Smith, 1973; Thorne, 1975), the break-up of 
Gondwanaland started too early to be relevant to any angiosperm distribution. To others 
(including myself), the distribution patterns that are found at present are so consistent with 
the plate tectonics theory that they must have been achieved with the help of continental 
drift. Such workers are thus forced to adopt the hypothesis that the early history of the 
angiosperms extends back to well before the Cretaceous, but that angiosperm fossils from 
that period have not been found or, more likely, have not been recognised. 

As the distribution of Hypericum seems to be interpretable only in terms of the plate 
tectonics theory, therefore, we are forced to postulate that the genus originated before direct 
land connections between Africa and the other parts of Gondwanaland were broken or, at 
least, before the areas of water between these continents became uncrossable by seeds or 
other diaspores. 


(b) Sect. 1. Campylosporus and the Macaronesian- Mediterranean area 


Whatever theory of plate tectonics that one adopts to reconstruct the distribution of 
continents before the break-up of Gondwanaland, the regions indicated in Figures 60 and 61 
are brought together round Africa (Fig. 73a, b). 

The least controversial distributional connections are those between Socotra and north- 
eastern tropical Africa, on the one hand, and Macaronesia and the Mediterranean area on 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 209 


the other (sects 2, 21-25, 27) (Figs 60, 61: tracks v, vii a). The disjunctions along these tracks 
are clearly due to the alpine orogeny and the formation of the Saharan desert belt. The north- 
ward migration of Africa resulted in the establishment of dry-land connections between 
Africa and Europe during the Cretaceous (between 148 and 80 m.y. BP) (Raven & Axelrod, 
1974; Raven, 1979). Whether north Africa and Macaronesia were part of Gondwanaland 
(Smith, Briden & Drewry, 1973) (Fig. 73a) or Atlantica (Melville, 1973 & ined.) (Fig. 73b) 
does not affect the issue. Sect. 23. Triadenioides will have had a dry-land link between 
Socotra and the Levant, and will have been able to reach Anatolia by the Eocene (Raven, 
1979): and the species of sects 24. Heterophylla and 25. Adenotrias are scattered round 
Morocco and the Mediterranean, mostly near the coast, as a result, presumably, of migration 
across the Saharan region by their ancestors. Of the Macaronesian sections, sect. 21. Webbia 
is a typical relict group with derivative relations in the Mediterranean (sect. 22. Arthro- 
phyllum). Sect. 27. Adenosepalum, in contrast, has primary derivative links with tropical 
Africa and only secondary ones across the Saharo-Sindian desert to (i) the Mediterranean 
and western Europe (including sect. 28. Elodes) and (11) the Himalayas and south-east Asia. 


(c) Sect. 20. Myriandra 


The disjunction between north-eastern Africa and the south-eastern United States (Figs 60, 
61: track vii b) indicates that the ancestors of sect. Myriandra must have crossed from Africa 
to North America before the Atlantic Ocean was formed, i.e. in the Jurassic (180 m.y. B.P.), 
or soon after its initiation. This long period of separation is consistent with the considerable 
morphological discontinuity between sects Campylosporus and Myriandra. Subsequent 
migrations within sect. Myriandra have no doubt been mostly over land, even to Cuba, 
which was probably reached via Mexico and Belize as well as from Florida (cf. Rosen, 1976, 
for links between Mexico and the Greater Antilles). 


(d) Sect. 3. Ascyreia and its derivatives 


The close resemblance between Hypericum revolutum subsp. keniense (sect. Campylo- 
sporus) and H. mysurense of south India and Sri Lanka (sect. Ascyreia) supports the theory 
that the latter’s ancestors reached the Indian plate before it broke away from Africa in the 
Mid-Cretaceous (c. 90-100 m.y. B.P.) (Raven, 1979). After the Indian plate abutted on Asia 
(45 m.y. B.P.), the species of sect. Ascyreia could have migrated across land to Europe, China 
and Malesia. The gap in distribution between the Himalayas and Anatolia (H. calycinum), as 
well as those between the Himalayas and the Caucasus (H. xylosteifolum— 
sect. 6. Inodorum) and Macaronesia (H. grandifolium—sect. 5. Androsaemum), indicates 
that an early westward trend was established by species of mesophytic habitats. These species 
are related respectively to H. oblongifolium (H. calycinum) and H. griffithii (H. grandi- 
folium, H. xylosteifolium). Another, possibly later, westward trend comprising species 
related to H. cordifolium and H. podocarpoides apparently gave rise to more xerophytic 
species, the progenitors of the Olympia group (sects 10-16) and the Hirtella group (sects 
17-19). In both these groups there is a gap between the Himalayan relatives and the 
Anatolian centre of variation—west Anatolia for the former, east Anatolia for the latter. 

The ancestors of sect. 4. Takasagoya probably reached Taiwan before it was separated 
from the Asian mainland, but whether the further migration to Luzon was over land or not is 
unclear. 

The other eastward development from sect. Ascyreia is that of sect. 7. Roscyna and its 
derivatives, sects 8-9a. The morphological gap between sects 3 and 7 is small, as is the 
geographical one (in China) between the areas of H. monogynum and H. elatoides. On the 
other hand, the whole of the central Asian plateau and desert area lies between the central 
and west Chinese H. elatoides group (H. elatoides, H. pedunculatum, H. przewalskii) and the 
west Transcaucasian H. bupleuroides (sect. 8. Bupleuroides), which suggests that the west- 
ward migration of the ancestor of H. bupleuroides occurred before the elevation of the 
Tibetan plateau. 


210 N. K. B. ROBSON 


Fig. 73 Two reconstructions of Gondwanaland and the north Atlantic area in the Triassic; (a) 
according to Smith, Briden & Drewry (1973); (b) according to Melville (ined.). In both 
reconstructions, solid lines indicate present shores, interrupted lines boundaries now on land, 
and dotted lines continental edges or fragments now submerged. 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 


211 


212 N. K. B. ROBSON 


Within one species of sect. Roscyna, H. ascyron, there is a wide disjunction between the 
Asian and eastern North American parts of its distribution. No evidence has been found to 
suggest that this species ever occurred in Europe,* and so the distributional links are likely to 
have been trans-Pacific. This hypothesis is made more likely when the trans-Pacific link 
between Asiatic forms of H. ascyron and the derivative Californian H. concinnum (sect. 9a. 
Concinna) are taken into consideration. The Bering bridge would seem to have been the 
route by which these species reached North America, if one accepts current orthodox views 
of plate tectonics in the Pacific area (cf. Raven & Axelrod, 1974). Even if Melville (1981) is 
correct and a Pacific continent did exist, it could not have provided a land-bridge directly 
from China to America for these species. By the time that the ocean gap between Asia and 
the western part of ‘Pacifica’ had narrowed sufficiently to allow the progenitors of sect. 
Roscyna to migrate there, the western (Chinese) and eastern (western North American) parts 
of that hypothetical continent would have been separated by a wide stretch of ocean. 

From similar considerations, it seems likely that the trans-Pacific link in sect. 9. 
Hypericum also indicates a Bering bridge migration in this section. The present-day nearest 
relatives are respectively in central China (H. pedunculatum, sect. Roscyna) and Mexico (see 
p. 192); but despite the more southerly distribution of the primitive species of American 
sect. Hypericum, there is no evidence of a southern origin of this group, i.e. of a northward 
divergence from ‘Pacifica’. Other genera with east Asian/Mexican disjunct areas are dis- 
cussed by Styer & Stern (1980), who suggest that they may all have had Tertiary pan-boreal 
distributions. 

To the westward, the H. maculatum—H. tetrapterum group reached western Europe and 
north Africa, apparently via a northern route and at a time when the Azores (H. undulatum) 
and the Faer6des (H. maculatum) were still reachable by land. H. perforatum penetrated west 
as far as the Canaries, south to Jebel Marra (Sudan Rep.) and east to China, apparently by 
natural means;} but its occurrence in Japan may not be due to natural spread, and its 
presence in North and South America, South Africa and Australasia is certainly the result of 
introduction. 


(e) Sects 1. Campylosporus and 26. Humifusoideum 


The first part of track iii (Figs 60, 61) has been considered in an earlier paper (Robson, 1979), 
where I discussed the close relationship between the Hypericum revolutum (African main- 
land) and H. lanceolatum (Mascarene Is.) groups, and the apparent parallel evolution that 
they have undergone. 

It was there concluded that they diverged from an ancestral species that probably occupied 
a single area in the east Africa-Madagascar-Mascarene region in the Cretaceous era, before 
these land blocks were separated by rift and drift. Although I accepted an original northern 
position for Madagascar in this paper (cf. Wild, 1975) (Fig. 73a), the evidence would fit in 
equally well (or better) with a southern position beside Mozambique (Fig. 73b). In addition, 
a southern position would make the distribution of sect. 26. Humifusoideum easier to inter- 
pret (cf. Melville, 1975). 

Sect. Humifusoideum has its primitive group of species (five in number—Robson, 1973a, 
1974) in the mountains of New Guinea. The most primitive of these, H. sewense, is most 
similar to H. lanceolatum (Réunion, Comoro Is.) and H. madagascariense (Madagascar). We 
can thus postulate a migration from the Madagascar-Mascarene region to Australia, with 
subsequent migration to New Guinea (Fig. 73b). Van Steenis (1979) explains how, in the 
Mid-Tertiary period, the Australian plate is thought to have made contact with east Malesia. 


*Reid & Reid (1915) record a seed from Pliocene deposits near the Dutch-German frontier as ‘Hypericum cf. 
ascyron’. A study of the specimen leads me to doubt if it belongs to Hypericum at all. It certainly cannot be regarded 
as good evidence for the occurrence of H. ascyron in Europe. 


+Wickens (1976), however, suggests that it may have been introduced to Jebel Marra by man. 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 213 


At the same time the previously archipelagic New Guinea region gradually emerged above 
the sea to form a large continuous land mass. The luxuriance of the subsequent influx of the 
Malesian element of the flora would have overwhelmed the Australian-Gondwanean 
element, suffocating it or allowing it to survive mainly in the recently formed mountains. 
The species of sect. Humifusoideum would then probably have died out in Australia as a 
result of the ensuing period of drought (Kemp, 1978). 

The diversification of sect. Humifusoideum, however, must have taken place before direct 
or indirect contact between Africa and Australia was lost in the Cretaceous period (Raven & 
Axelrod, 1974). The most primitive African species, H. natalense, is very similar to the 
mostly widely dispersed New Guinea species, H. papuanum; so it seems that representatives 
of sect. Humifusoideum must have reached the African mainland after speciation had 
occurred in Australia. H. natalense is confined to extreme south-eastern Africa, but the 
derivative H. wilmsii has a disjunct distribution in South Africa, the eastern border of 
Zimbabwe and Madagascar. Therefore, unless Madagascar was reached by long-distance 
dispersal (for which there is no evidence), this is another reason for putting the original 
spread of the African species back to the era before contact between Africa, Madagascar and 
Australia was finally broken, about 140 m.y. B.P. (Melville, 1975). India may or may not 
have been part of the relevant area of contact (Fig. 73b). At any rate, this section does not 
occur in India or Sri Lanka today; and the occurrence in that region of sect. 3. Ascyreia, on 
track iv, would suggest that sect. Humifusoideum (on track iii) took a more southerly course. 
These considerations all favour a reconstruction of Gondwanaland (such as that in Fig. 73b) 
where Madagascar has a southern position and is close to Australia, rather than one (such as 
in Fig. 73a) where it has a northern position and Australia is separated from Africa by 
Antarctica. 

The above reconstruction of events does not, however, account for the occurrence of H. 
pulogense in the mountains of Luzon (Philippines) and of H. beccarii in scattered regions of 
Sumatra and west Java. These species are advanced relative to the New Guinea ones and well 
separated from them morphologically as well as geographically. It is quite possible, then, 
that they have reached their present areas of distribution gradually over land, possibly when 
the Sunda shelf was largely dry land during the Glacial Epoch (van Steenis, 1979). H. 
pulogense is on van Steenis’s New Guinea track (New Guinea to Celebes and the 
Philippines), one of the dispersal routes that he distinguished when discussing the origin of 
the Malaysian mountain flora (van Steenis, 1934). The occurrence of H. beccarii, on 
Sumatra and Java, on the other hand, is less easily explained. Perhaps it reached these islands 
via Sulawesi. 


(f) Sect. 29. Brathys 


Similar arguments to those that have been adduced to explain the occurrence of sect. 20. 
Myriandra in Belize and Cuba as well as in eastern North America can explain the presence 
of the primitive species of sect. 29. Brathys (Hypericum styphelioides) in Belize and Cuba 
(Figs 60, 61: track i). The more southerly distribution of the primitive members of the section 
can be correlated with the more southerly distribution in Africa of its nearest relative (H. 
revolutum subsp. keniense); but too much emphasis should not be placed on this argument. 
It is more important to emphasise the ‘vicariance model of Caribbean biogeography’ 
described by Rosen (1976), whereby the original link between North and South America, in 
the late Mesozoic, was via the proto-Antilles, from Yucatan via Cuba, Hispaniola, Puerto 
Rica and the Lesser Antilles, which at that time are thought to have formed a connecting 
archipelago. When North and South America began to move westward from Africa, the 
proto-Antilles archipelago moved less quickly and was deformed to an arc. Subsequently, in 
the Middle to Late Tertiary, another volcanic archipelago was formed between Mexico and 
Colombia, and this consolidated to produce the present Central American isthmus. 

This reconstruction implies that the ancestors of sect. Brathys reached America from 
Africa in one of two ways. (i) They may have arrived first in northern South America, 


214 N. K. B. ROBSON 


speciated there and migrated early northward along the proto-Antilles in two or three waves, 
the first of which (H. styphelioides) reached Belize and Cuba whereas the second (H. 
millefolium) and third (?) (H. pycnophyllum) only reached Hispaniola. (ii) Alternatively, 
they may have arrived first in the extreme south of the North American plate, in Belize and 
Cuba, migrated southward along the proto-Antilles, and subsequently retraced their tracks as 
far as Hispaniola. Of these two possibilities the first involves the more direct routes and is 
therefore the preferred hypothesis. 

Whichever way the original migrations occurred, there has clearly been much subsequent 
movement to-and-fro along the Andean range; and the Costa Rican mountains have been 
reached from South America by three species that are not closely related to one another (H. 
stenopetalum, H. caracasanum, H. strictum sens. lat.). In addition, the presence of an 
endemic species on Mt Roraima (H. roraimense) may well date from an earlier period in the 
evolution of sect. Brathys, i.e. before the Andean range was formed, as this mountain 1s part 
of the ancient Guyana Highland (Maguire, 1970). 


(g) Sect. 30. Spachium 


It has been possible to explain the distribution of previous sections without recourse to a 
hypothesis of long-distance dispersal. This is not to deny, of course, that such dispersal has 
occurred in them, but only to point out that it need not have done so. In sect. 30. Spachium, 
on the other hand, we are faced with a section of which the current distribution 1s apparently 
partly inexplicable except in terms of long-distance dispersal. This exceptional behaviour, 
moreover, becomes more plausible if one considers the habit and habitats of the relevant 
species. They are all annual or shortly perennial herbs of wet places (ditches, river banks, 
lake margins, rice fields), with seeds that are relatively small for Hypericum. These 
considerations make them eminently suitable for dispersal in mud attached to the feet of 
birds, although there is no direct evidence for this (see p. 124). 

The differences between H. revolutum subsp. keniense and H. rigidum, the most primitive 
species in sect. Spachium, are both considerable and consistent with the hypothesis that the 
ancestors of these taxa diverged before South America broke away from Africa in the 
Cretaceous. H. rigidum, in southern Brazil, occupies one of the most ancient continental 
areas in South America. This hypothesis is also supported by the existence of a closely 
related and taxonomically difficult group immediately related to (derived from ?) H. rigidum 
in the area to the south of Amazonia, stretching from the Brazilian coastal mountains to the 
foothills of the Andes in Bolivia (Fig. 71). This group of species (the H. brasiliense—H. 
connatum group) must have evolved over a long period in order to produce 22 species that 
are morphologically so varied. Yet, the fact that the gaps in this variation are mostly small 
suggests that confinement to the region between the hot wet Amazon and the cool dry 
pampas has resulted in repeated partial isolations and hybridisation, again over a long 
period. 

The disjunction between the H. denticulatum group of south-eastern U.S.A. (H. denti- 
culatum sens. lat., H. setosum and H. cumulicola) and its nearest relatives, in southern Brazil 
(H. cordiforme and H. ternatum) is difficult to interpret other than in terms of ancient long- 
distance dispersal. The morphological gap between H. denticulatum and the Brazilian 
species is not great. 

The rest of this large section is apparently derived from near H. chamaemyrtus, a relict 
species of the Colombian Andes that occurs in distinct forms in two relatively widely 
separated localities (provinces Santander and Cundinamarca). H. chamaemyrtus appears to 
be directly but rather distantly related to H. rigidum; but the areas of these species are 
separated by the whole width of the Amazonian rainforest, a region that has been drier than 
it is at present (Prance, 1978). It seems possible, then, that the ancestors of H. chamaemyrtus 
migrated (i) directly over this area or (ii) via the (new) Andes, or that (i11) the ancestors of H. 
rigidum/H. chamaemyrtus occupied some intermediate area. The Andean route would seem 
to be the most likely of these possibilities (cf. Simpson, 1975). 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 215 


Not only is H. chamaemyrtus taxonomically isolated, but it is the apparent origin of 
several separate evolutionary lines, all of which show more-or-less wide disjunctions (Figs 
71, 72). In one line, in which the basic chromosome number (x) is 12, the gap between 
Colombia and the Honduras Republic (H. gnidioides) is consistent with a migration through 
the proto-Antilles to Central America, whence there is a secondary radiation involving 
further disjunctions. There are, and were, direct overland connections south to Panama and 
north to Mexico for H. gnidioides and some derivative species, and the H. dichotomum 
group (Greater Antilles) could have diverged from Central America back along the proto- 
Antilles; but how H. eastwoodianum reached the Revillagigedo Islands is not clear, as this 
species also occurs in western Mexico (Jalisco, Zacatecas). H. drummondii* and H. 
gentianioides, of the eastern United States, may well have reached their present area over 
land from Mexico; but the remaining species in this line, H. piriai and H. hilaireanum of 
southern Brazil and Uruguay, have links across Amazonia again. These species are morpho- 
logically isolated in the section and appear to represent an early southward migration from 
the H. chamaemyrtus/H. gnidioides line. 

H. pauciflorum is also closely related to H. chamaemyrtus. It and its immediate 
descendents have spread north and south along Central America, but from a more northern 
centre (Mexico), north to Texas and south to Panama. 

There is then a disjunction northwards to the H. majus group, all of which have n = 8, and 
all except H. anagalloides (western United States and Canada) have apparently indulged in 
long-distance dispersal. Thus H. majus itself (northern United States and Canada) also 
occurs in France and Germany and has recently been found in Japan (Hokkaido) (Hara, 
Kurosawa & Inteishi 13.ix.1974, TI), H. canadense (north-eastern United States and 
Canada) also occurs in western Ireland and the Netherlands, and H. gymnanthum (south- 
eastern United States) also occurs in Guatemala and Poland. H. gymnanthum, in turn, has 
given rise to two widespread groups: 

(i) H. mutilum (possibly including H. boreale (Gillett, 1979)) is native to the eastern and 
central United States and Canada; but it has been found in many other parts of the world, 
viz. Mexico, Honduras Republic, Cuba, Hispaniola, Colombia, Ecuador, Brazil, 
Paraguay, Uruguay, Argentina, Peru, Hawaii, the Azores, France, Italy, Poland and 
Transcaucasia. Some of these introductions have probably been recent (e.g. to Italy and 
Hawaii); others must be of considerable antiquity, because H. pleiostylum (eastern 
Brazil) and H. arenarioides (Cuba) are both apparently derivatives of H. mutilum. 

(ii) H. japonicum, a widespread and variable ricefield weed of east Asia, has many wide 
disjunctions in its total area (see p. 206 and Robson, 1973a). Like the other species just 
mentioned, its distribution seems capable of interpretation only in terms of long-distance 
dispersal (by birds ?). The original dispersal from Texas (?) to south-east Asia, like the 
long-distance ones of H. mutilum, must have been of considerable antiquity for the 
species to have achieved its present wide distribution and varied morphology. 

The very close morphological links between Mexican H. pauciflorum and a related species 
in the Colombian Andes, on the one hand, and primitive forms of H. /alandii in Natal and 
the Transvaal, on the other, may be the result of direct intercontinental migration. If, 
however, one considers (i) the apparently long prior evolution of sect. Spachium in South 
America, (ii) the northward African trends in the H. /alandii group and (iii) the apparent 
behaviour of some other herbaceous members of this section, then ancient long-distance 
dispersal from northern South America or Central America seems a much more probable 
explanation of the present disjunction. The presence of H. /alandii and the derivative H. 
perrieri group in Madagascar would then probably be best explained by long-distance 
dispersal from the African mainland. The other descendents of H. lalandii (H. humbertii, H. 
scioanum and H. oligandrum) are all confined to the African mainland. 


*Reid (1923) identified Pliocene fossil seeds from Pont-de-Gail, Cantal, France as belonging to Hypericum (H. 
cantalense E. M. Reid) and compared them, and other Pliocene seeds from the Dutch—German frontier at Reuver, 
with H. drummondii. The comparison does not seem very apt to me, and I doubt very much if the seeds do belong to 
Hypericum. 


216 N. K. B. ROBSON 


Finally, the distribution of the H. silenoides group is problematic. This species is clearly 
derived from the Colombian H. chamaemyrtus and has primitive forms in two areas, the 
Ecuador-Peru border (H. silenioides sensu stricto) and northern coastal Chile (H. paposum I. 
M. Johnston). From the former area there are gradual morphological trends along the Andes 
northward to Venezuela and southward to northern Argentina, as well as a fairly wide 
geographical disjunction between Bolivian forms of H. silenoides and H. caespitosum of 
central Chile. The nearest relatives of ‘H. paposum’, on the other hand, are separated from it 
by the width of the Pacific Ocean. They are the most primitive forms of H. gramineum, 
which occur in New Zealand and south-eastern Australia. This distribution at first suggests 
an Antarctic distribution pattern; but, whether Magellania was part of South America when 
it split from Africa (e.g. Smith, Briden & Drewry, 1973) (Fig. 73a) or not (Melville, 1966) 
(Fig. 73b), it would not have been possible for the ancestors of H. gramineum to have 
reached there from northern South America at a time when the Antarctic migratory path was 
continuous. Thus, here again we seem to be faced with an ancient long-distance dispersal, 
this time between western South America and Australia/New Zealand. Whether or not this 
disjunction was the result of long-distance dispersal, however, there seems to be no doubt 
that the subsequent spread of H. gramineum (from (i) New Zealand to New Caledonia, New 
Guinea and Taiwan; (ii) Australia to Vietnam and the eastern Himalaya) was by this means 
(cf. Robson, 1973a: fig. 3). 


11. Summary 


1. Hypericum steyermarkii Standley is more primitive than other species of Hypericum in having 5 
sterile structures representing the antisepalous whorl of stamen fascicles. In other characters (e.g. 
leaf venation) it is more advanced. It has therefore been removed to a separate genus, Santomasia 
gen. nov. 

2. The Bonnetioideae sensu Maguire (1972), i.e. including the Kielmeyeroideae, should be included 
in the Guttiferae. Although this subfamily is more primitive in most characters than the Hyperi- 
coideae, it is not directly ancestral to the latter. Indeed, the three tribes of the Hypericoideae seem 
to show relationship to different parts of the Bonnetioideae. 

3. Correlation of trends, rather than of characters, gives a more reliable indication of the course of 
evolution in a given taxon; and the mapping of correlated trends in modern taxa often suggests the 
direction in which the evolution has occurred. 

4. A diagram showing the apparent relationships of the sections of Hypericum can be used to show the 
distribution of characters throughout the genus. 

5. By back-projection of trends, it is possible to visualise the primitive Hypericum as being a tall shrub 
or small tree with 4-lined stem-internodes, entire sessile leaves having (i) a midrib and + numerous 
veins that dichotomise at or near the base and then run parallel and (ii) pellucid, possibly only 
punctiform glands. The flowers were large, solitary and terminal, and bracteoles were not differen- 
tiated. The floral whorls were all 5-merous, the outer (antisepalous) whorl of stamen fascicles being 
absent. The sepals were completely free and, like the petals, entire with linear pellucid glands 
between the parallel veins. Dark glands were probably absent from all organs. The 5 antipetalous 
stamen-fascicles each had numerous stamens with filaments distinct from near the base. The pollen 
grains were prolate-spheroidal, with the endoaperture porate or almost so and the ornamentation 
possibly a tectum perforatum with tectal perforations grouped together. The ovary had 5 placentae, 
which formed a central column but were not markedly axile (i.e. not closely united), and was 
surmounted by 5 free but + appressed styles. The seeds were numerous, probably narrowly 
unilaterally winged, and with a reticulate testa. H. bequaertii resembles this hypothetical ancestor 
closely. 

6. From this hypothetical (but probably actual) species, evolutionary trends in various directions have 
resulted in: Shrubs or-herbs with terete stem-internodes, leaves sometimes sub-petiolate, pinnate to 
reticulate-veined, rarely gland-fringed or scale-like, with dark punctiform glands; inflorescences of 
many small flowers, primarily cymose but occasionally partially racemose, with differentiated 
bracteoles; meiomerous floral whorls (the outer ones sometimes reduced to 4-mery, the gynoecium 
to 2-mery); sepals + united, with punctiform pellucid or linear dark laminar glands and variously 
glandular or eglandular margin; petals remaining free, but otherwise varying like the sepals except 


16: 


|e 
18. 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 217 


for the rare appearance of a ligule; stamen-fascicles united in various ways, sometimes with the 
disappearance of the common ‘trunk’, or rarely with the union of the filaments becoming more 
pronounced, and with a reduction in the number of stamens per fascicle; pollen becoming prolate 
with the endoaperture a lalongate colpus and the ornamentation reticulate with tectal perforations 
regularly spaced; the ovary with completely axile or completely parietal placentation and styles 
that are + completely united or completely free and divergent; seeds without a wing or carina, the 
testa becoming ribbed and striate or papillose. 


. The primitive Hypericum flower is a brush blossom, open-pollinated by less-specialised insects 


(e.g. Syrphidae), which are attracted probably wholly visually by the ‘target’ appearance of the 
open flower (bee-yellow centre, bee-purple surroundings). In two separate sections, a change to 
specialised insect-pollination has been accompanied by the evolution of a pseudo-tubular flower 
and other modifications including, in sect. Adenotrias, dimorphic heterostyly. In both sections, 
members of the antisepalous stamen-fascicle have reappeared, acting like lodicules in opening the 
flower. 


. Seeds of Hypericum are small, but they appear to be dispersed largely by gravity. No long-distance 


dispersal has been demonstrated, although circumstantial evidence for this in sect. Spachium is 
strong. 


. Normally, Hypericum has a Polygonum-type embryo sac and the embryogeny is Solanad. 


Recorded abnormalities include polyembryony, pseudo-polyembryony (concrescence of ovules) 
and apospory. The germination is epigeal. The floral whorls initiate development in order 
centripetally, the sepals arising quincuncially, the petals simultaneously. The stamen-fascicles 
develop more quickly at first than do the petals, and the individual stamens arise in centrifugal 
succession on each fascicle primordium. 


. Variation in the vascular structure of the torus parallels that established from morphological 


studies, being interpretable in terms of a basic pattern of regularly pentamerous whorls, the three 
outer ones each with unilacunar traces. This variation enables one to discriminate between com- 
plete loss of an organ and ‘loss’ by union (i.e. loss of individuality). From a comparison of sections 
of the stele at toral level, it is possible to show that it has undergone both vertical and horizontal 
contractions. 


. Floral morphology and vasculature in Hypericum are most easily interpreted in terms of the 


Gonophyll Theory (Melville, 1962, 1963). 


. Variation in the essential oils of Hypericum is not of great taxonomic importance; but the presence 


and distribution of hypericin and pseudo-hypericin is characteristic and sometimes diagnostic of 
sections or groups of sections. 


. Counts of Hypericum chromosomes published since 1968 have confirmed the general picture 


described by Robson & Adams (1968), viz. a reduction of the basic number in different evolution- 
ary lines from 12 to 6, with the occurrence of tetraploidy on most basic numbers. Higher poly- 
ploidy seems to be associated with reproductive abnormality. One secondary basic number (21) has 
been reported. 


. Natural hybridisation in Hypericum is uncommon. It has been shown to occur in sects 5. Andro- 


saemum and 9. Hypericum, and its presence is suspected in sects 20. Myriandra and 30. Spachium. 


. Artificial hybrids have been confined mainly to sects 3. Ascyreia and 9. Hypericum. In the former 


section H. x ‘Hidcote’ is thought on circumstantial evidence to have resulted from crossing H. 
calycinum and H. beanii ‘Gold Cup’. 

In general: Crosses in Hypericum (1) often result in variegation, (i1) do not proceed beyond the 
seedling stage where different chromosome numbers are involved, and (iii) rarely yield seeds when 
‘wide’. When H. perforatum (4x) crosses with diploid species of sect. Hypericum, two forms of 
hybrid (3x, 5x) are produced. 

H. concinnum Bentham is related to a different part of sect. 7. Roscyna from the other species in 
sect. 9. Hypericum. It has therefore been placed in a separate section, 9a. Concinna sect. nov. 

Of the three tribes of the Hypericoideae: the Vismieae appears to be primitively South American 
and to have spread to Central America, Africa and Madagascar; the Cratoxyleae is primitively 
Malesian (or perhaps Madagascan) and has spread to eastern Asia and thence to (i) eastern North 
America and (ii) Mexico/Guatemala; and the Hypericeae is primitively African (Hypericum) and 
Central American (Santomasia). 


. Morphological and distributional trends in Hypericum indicate that the genus probably originated 


in tropical Africa and that it has spread initially to South America (sects 30. Spachium and 
229. Brathys), North America (sect. 20. Myriandra), north Africa, the Mediterranean and Socotra 
(sects 2. Psorophytum, 21. Webbia, 23. Triadenioides, 27. Adenosepalum), south India and Sri 


218 N. K. B. ROBSON | 


Lanka (sect. 3. Ascyreia) and New Guinea (via Australia ?) (sect. 26. Humifusoideum). There is no 
evidence of long-distance dispersal at this stage. Indeed, all these migrations (of tree and shrub 
species) could have taken place over dry land if Hypericum originated and spread initially before 
the break-up of Gondwanaland; and the ‘radiation’ from Africa is best interpreted in terms of such 
a theory. 

20. Subsequent evolution of herbaceous species in sect. Spachium has resulted in distributions strongly 
suggestive of long-distance dispersal, possibly by wading birds and waterfowl. Despite the 
evolution of many perennial and a few annual herbs in other sections, the evidence for long- 
distance dispersal elsewhere in the genus is almost absent, most distributions being interpretable in 
terms of over-land migration. Only in sect. Hypericum (mountains of Sumatra and Borneo) does 
such a migration seem unlikely. 


12. Acknowledgements 


I am very grateful to Dr Giles Clarke, for making a study of Hypericum pollen; to Dr Mary Gibby, for 
her studies of Hypericum chromosomes, and to Mme Claude Reynaud for information on the same 
subject; to Peter Dummer and Donald Walker, for hybridising various species of Hypericum; to Roy 
Lancaster, Tony Schilling and Allen Paterson for cultivating various Hypericum species and providing 
material and data; to Mrs Isobyl La Croix, for providing seeds of Hypericum revolutum and data on its 
growth; to Dr Bassett Maguire, for providing material of his Bonnetioid genera, as well as for help in 
other ways; to Dr Susan Jones, for data on Garcinia and other non-Hypericoid genera; and to these 
and many other colleagues and friends, particularly Dr Ronald Melville, for helpful discussions. 

With regard to this paper, I very gratefully acknowledge help from Mrs Margaret Tebbs, for her 
original drawings and diagrams, as well as for improving my own drawings of floral vasculature; to Miss 
Marian Short, for drawing the distribution data on the maps and taking the S.E.M. photographs; to Bob 
Press for drawing some of the outline maps; and to Miss Loveday Hosking for typing the manuscript. 


13. References 


Adams, W. P. 1962a. Studies in the Guttiferae. I. A synopsis of Hypericum section Myriandra. Contr. 
Gray Herb. Harv. Univ. 189: \-51. 

— 1962b. Studies in the Guttiferae. I]. Taxonomic and distributional observations on North 
American taxa. Rhodora 64 : 231-242. 

— 1972. Studies in the Guttiferae. III. An evaluation of some putative garden hybrids in Hypericum 
sect. Myriandra. Rhodora 74 : 276-282. 

André, E. 1889. Hypericum Moserianum. Revue hort. 61 : 463-464, t. 116, 117. 

Ashlock, P. D. 1974. The use of cladistics. Ann. Rev. Ecol. Syst. 5 : 81-99. 

Baas, P. 1970. Anatomical contributions to plant taxonomy I. Floral and vegetative anatomy of Eliaea 
from Madagascar and Cratoxylum from Indo-Malesia (Guttiferae). Blumea 18 : 369-391. 

Ball, I. R. 1976. Nature and formulation of biogeographical hypotheses. Syst. Zool. 24 : 407-430. 

Bamps, P. 1966. Notes sur les Guttiferae d’Afrique tropicale. Bull. Jard. bot. Etat Brux. 36 : 425-459. 

1971. Notes sur les Guttiferae d’ Afrique tropicale. Bull. Jard. bot. nat. Belgique 41 : 427-444. 

Baretta-Kuipers, J. 1976. Comparative wood anatomy of Bonnetiaceae, Theaceae and Guttiferae. 
Leiden botl Series 3 : 76-101. 

Bawa, S. B. 1970. Theaceae. Jn Symposium on comparative embryology of Angiosperms. Bull. Indian 
natl Sci. Acad. 41 : 75-77. 

Bean, W. J. 1973. Trees and Shrubs Hardy in the British Isles 2 : 405-424. 8th ed. London. 

Bell, C. R. 1965. Documented plant chromosome numbers, 65. Sida 2 : 168-170. 

Bentham, G. 1862. Hypericineae, Guttiferae. Jn G. Bentham & J. D. Hooker, Genera Plantarum 
1: 163-177, London. 

Borgen, L. 1969. Chromosome numbers of vascular plants from the Canary Islands, with special 
reference to the occurrence of polyploidy. Nytt Mag. Bot. 16 : 81-121. 

Brandza, G. 1908. Recherches anatomiques sur la germination des Hypéricacées et des Guttiféres. Ann. 
Sci. nat. (Bot.) IX, 8 : 221-300. 

Breedlove, D. E. & McClintock, E. 1976. Thornea (Hypericaceae), a new genus from Mexico and 
Guatemala. Madrorio 23 : 368-373. 

Breindl, M. 1934. Zur Kenntnis der Baumechanik des Bliitenkelches der Dikotylen. Bot. Arch. 


36 : 191-268. 
- ” 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 219 


Brenan, J. P. M. 1945. Notes on the flora of Oxfordshire and Berkshire. Rep. botl Exch. Cl. Br. Isles 
12 : 781-802. 

Brown, E. G. S. 1935. The floral mechanism of Saurauja subspinosa Anth. Trans. bot. Soc. Edinb. 
31 : 485-497. 

Bugnicourt, M. 1970. Polyembryonie chez l’Hypericum tetrapterum Fries (= Hypericum acutum 
Moench). Bull. Soc. bot. Fr. 116 : 495-504. 

— 197la. Embryogénie des Hypericacées. Developpement de |’embryon chez |’Hypericum tetrap- 
terum Fries. Bull. Soc. bot. Fr. 118 : 319-334. 

— 1971b. Polyembryonie resultant de la concrescence de deux ou de plusieurs ovules chez quelques 
espéces du genre Hypericum. Ann. Univ. ARERS, Reims 9 : 306-310. 

Caraz-Billat, M. M. 1968. Contribution a l'étude botanique d’'Hypericum nummularium de la Grande 
Chartreuse. Thesis, Sect. Pharmacie, Univ. Lyon. 

Chakravarty, H. L. 1955. Morphology of the staminate flowers of the Cucurbitaceae with special 
reference to the evolution of the stamens. Bull. bot. Soc. Bengal8 : 186-215. 

Choisy, J. D. 1821. Prodromus d'une monographie de la famille des Hypéricinées. Geneva & Paris. 

— 1824. Hypericineae. Jn A. P. de Candolle, Prodromus systematis naturalis Regni vegetabilis 
1 : 541-556. Paris. 

Clarke, G. C. S. 1975. Irregular pollen grains in some Hypericum species. Grana 15 : 117-125. 

— 1976. The northwest European pollen flora, 7. Guttiferae. Rev. Palaeobot. Palynol. 21 : 125-142. 

Clos, M. D. 1868. Des glandes dans le genre Hypericum. Mém. Acad. Sci. Toulouse 6 : 257-266. 

Connor, H. E. 1977. The Poisonous Plants in New Zealand. 2nd ed. Wellington. 

Contandriopoulos, J. & Lanzalavi, M. 1968. Contribution a |’étude cytotaxinomique des Hypericum de 
Gréce. Bull. Soc. bot. France 115 : 5-14. 

Corner, E. J. H. 1946. Centrifugal stamens. J. Arn. Arbor. 27 : 423-437. 

— 1976. The Seeds of Dicotyledons. Cambridge. 

Correns, C. 1931. Hypericum perforatum status albomaculatus. Sber. preuss. Akad. Wiss., Phys. 
-Math. KI. 1931 : 209-215. 

Costa, M. J. F. 1904. Hypericum androsaemum L. Pharm. thesis, Univ. of Coimbra. 

Coutinho, M. C. P. 1950. Notas sobre a constituigao histo-anatomica das diversas espécies do genero 
Hypericum L. existentes na Serra do Gerés. Agron. lusit. 12 : 517-549. 

Croizat, L., Nelson, G. & Rosen, D. E. 1974. Centers of origin and related concepts. Syst. Zool. 
23 : 265-287. 

Cronquist, A. 1968. The Evolution and Classification of Flowering Plants. London. 

Culwell, D. E. 1970. A Taxonomic Study of the Section Hypericum in the Eastern United States. Ph.D. 
Thesis, Univ. of N. Carolina, Chapel Hill. (See Diss. Abstr. Int. B, 31 : 6451). 

Dahlgren, R. 1977. A commentary on a diagrammatic presentation of the Angiosperms in relation to 
the distribution of character states. Plant Syst. Evol., Suppl. 1 : 253-283. 

Davis, G. L. 1966. The Systematic Embryology of Angiosperms. New York. 

Dehay, J. 1972. Prospection caryologique en Brenne et Limousin. Bull. Soc. bot. France, 116 (97° sess. 
extraord.) : 69-74. 

Druce, G. C. 1923. Hypericum hirsutum L. In L. V. Lester Garland, Report of the distributor for 1922. 
Rep. botl Exch. Cl. Br. Isles 6 : 829. 

Dupuy, P. & Guédeés, M. 1975. Placentation and possible partial stachyospory in Hypericum sect. 
Eremanthe. Flora, Jena 164 : 37-49. 

Dyer, W. J. J. 1874. Hypericineae. Jn J. D. Hooker, Flora of British India 1 : 252-258. 

Eisner, J., Eisner, M. & Aneshansley, D. 1973. Ultraviolet patterns on rear of flower buds: basis of 
disparity of buds and blossoms. Proc. nat. Acad. Sci. U.S.A.70 : 1002-1004. 

Engler, A. 1925. Guttiferae. Jn A. Engler & K. Prantl, Die natiirlichen Pflanzenfamilien. 2nd ed. 
21: 154-237. 

Erdtman, G. 1971. Pollen Morphology and Plant Taxonomy. Angiosperms. 2nd ed. New York. 

Ewan, J. 1962. Synopsis of the South American species of Vismia (Guttiferae). Contr. U.S. nat. Herb. 
35 : 293-377. 

Faegri, K. & Iversen, J. 1975. Textbook of Pollen Analysis. 3rd ed. Kobenhavn. 

— & van der Pijl, L. 1971. The Principles of Pollination Ecology, 2nd ed. Oxford. 

Fahrenholtz, H. 1927. Uber Rassen- und Artkreuzungen in der Gattung Hypericum. In Anon. (ed.) 
Festschrift fiir Schauinsland. Bremen. 

Farris, J.S. 1970. Methods of computing Wagner trees. Syst. Zool. 19 : 83-92. 

Feradkova, V. 1972. Bemerkungen zur Flora der Umgebung der Stadt Hlohovec in der Slowakei III. 
Erganzungen. Acta Fac. Rerum nat. Univ. Comen. (Bot.) 20 : 149-158. 


220 N. K. B. ROBSON 


Fernald, M. L. 1947. Additions to and subtractions from the flora of Virginia. Rhodora 49 : 85-115, 
121-142, 145-159, 175-194. 

Fletcher, H. R. 1955. Award of Garden Merit—XC.427. Hypericum patulum ‘Hidcote’. J. R. hort. Soc. 
80 : 551-552, t. 129. 

Frohlich, A. 1915. Uber zwei der Steiermark eigentumliche Formen aus dem Verwandschaftkreis des 
Hypericum maculatum Cr. Mitt. Nat. Ver. Steiermark 51 : 216-245. 

— 1960. Krizenci trezslek (Hypericum) v C.S.R. (Zwei Hypericum-bastarden in der 
Tschechoslovakei). Preslia 32 : 97-99. 

Gadella, T. W. J. & Kliphuis, E. 1966. Chromosome numbers of flowering plants in the Netherlands. 
II. Proc. Koninkl. Akad. Wetens. C, 69 : 541-546. 

—— —— 1970. Cytotaxonomic investigations in some Angiosperms collected in the valley of Aosta 
and in the National Park ‘Gran Paradiso’. Caryologia 23 : 363-369. 

Gagnieu, A. & Wilhelm, J. P. 1965. Genre Hypericum [In A. Gagnieu, Les chromosomes dans la 
cellule, la plante, l’espéce]. Jn Anon. (Ed.) Travaux biologique dédit a Prof. Plantefol: 472-473. 
Paris. 

Gensini, F. 1967. Influenza della luce sulla germinazione di Hypericum perforatum L. G. bot. ital. 
100 : 299-300. 

Gillett, J. M. 1975. In A. Love (Ed.), IOPB chromosome number reports, L. Taxon 24 : 671-678. 

— 1979. New combinations in Hypericum, Triadenum, and Gentianopsis. Can. J. Bot. 57 : 185-186. 

Gliick, H. 1911. Biologische und morphologische Untersuchungen tiber Wasser- und Sumpfgewdachse 
3 : 38-41. Jena. 

Goebel, K. 1898-1901. Organographie der Pflanzen inbesondere der Archegoniaten und 
Samenpflanzen. Jena. 

Gogelein, A. J. F. 1967. A revision of the genus Cratoxylum BI. (Guttiferae). Blumea 15 : 453-475. 

Gorshkova, S. G. 1949. Guttiferae. Jn B. K. Shishkin & E. G. Bobrov (Eds) Flora U.R.S.S. 15 : 
210-258. 

Govindappa, D. A. 1956. Morphological study of Hypericum japonicum Thunb. J. Mysore Univ. 
15 : 69-79. 

Green, J. R. 1884. On the organs of secretion in the Hypericaceae. J. Linn. Soc. (Bot.) 20 : 451-464. 

Gunatilaka, A. A. L., Balasubramanian, S. & Kumar, V. 1979. 2,3-dimethoxyxanthone from 
Hypericum mysorense. Phytochem. 18 : 182-183. 

Hallé, F., Oldeman, R. A. A. & Tomlinson, P. B. 1978. Tropical Trees and Forests. An Architectural 
Analysis. Berlin, Heidelberg & New York. 

Hedberg, J. & Hedberg, O. 1977. Chromosome numbers of afroalpine and afromontane angiosperms. 
Bot. Notiser 130 : 1-24. 

Heel, W. A. van 1966. Morphology of the androecium in Malvales. Blumea 13 : 177-394. 

Heine, H. 1962. Les Millepertuis américains dans la flore d'Europe. Bauhinia 2 : 71-78. 

Hennig, W. 1966a. Phylogenetic systematics. Urbana. 

— 1966b. The Dipteran fauna of New Zealand as a problem in systematics and zoogeography. Pac. 
Insects Monogr. 9 : 1-81. [Transl. P. Wygodzinsky]. 

Henslow, G. 1890. On the vascular systems of floral organs, and their importance in the interpretation 
of the morphology of flowers. J. Linn. Soc. (Bot.) 28 : 151-197. 

Herbst, W. 1935. Uber Kreuzungen in der Gattung Hypericum, mit besonderer Beriicksichtigung der 
Buntblatterigkeit. Flora, Jena 129 : 235-259. 

Hickey, L. J. 1973. Architecture of dicotyledonous leaves. Amer. J. Bot. 60 : 17-33. 

— & Doyle, J. A. 1977. Early Cretaceous fossil evidence for angiosperm evolution. Bot. Rev. 
43 : 3-104. 

—— & Wolfe, J. A. 1975. The basis of angiosperm phylogeny: vegetative morphology. Ann. Mo. bot. 
Gdn 62 : 538-589. 

Hirmer, M. 1917. Beitraége zur Morphologie der polyandrischen Bliiten. Flora, Jena 110 : 140-192. 

Hoar, C.S. 1931. Meiosis in Hypericum punctatum. Bot. Gaz. 92 : 396-406. 

—— & Haertl, E. J. 1932. Meiosis in the genus Hypericum. Bot. Gaz. 93 : 197-204. 

Hochreutiner, B. P. G. 1918. La fonction ‘lodiculaire’ des corpuscles hypogynes chez les Guttiferes. C. 
r. Séanc. Soc. Phys. Hist. nat. Genéve 35 : 82-85. 

1919. La parenté des Guttiféres et des Hypericinées. C. r. Séanc. Soc. Phys. Hist. nat. Genéve 
36 : 26-27. 

Hofmeister, W. 1868. Handbuch der physiologen Botanik. 1 (2). Allgemeine Morphologie der 
Gewachse. Leipzig. 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 221 


Holm, T. 1903. Triadenum virginicum (L.) Rafin. A morphological and anatomical study. Am. J. Sci. 
IV, 16 : 269-276. 

Holub, J., Mesicek, J. & Javurkova, V. 1972. Annotated chromosome counts of Czechoslovak 
plants. Folia Geobot. Phytotax. 7 : 167-202. 

Hsu, C.-C. 1968. Preliminary chromosome studies on the vascular plants of Taiwan, II. Taiwania 
14: 11-27. 

Hughes, N. F. 1977. Palaeo-succession of earliest Angiosperm ancestors. Bot. Rev. 43 : 105-127. 

Hutchinson, J. 1959. The Families of Flowering Plants. I. Dicotyledons. 2nd ed. Oxford. 

Huyhn, K. L. 1965. Contribution a I’étude caryologique et embryologique des phanérogames du Pérou. 
Denkschr. schweiz. nat. Ges. 85 : 1-178. 

Isikawa, S. 1962. Light sensitivity against the germination. III. Studies in various partial processes in 
light-sensitive seeds. Jap. J. Bot. 18 : 105-132. 

Ivimey-Cook, R. B. 1963. Biological flora of the British Isles. Hypericum linarifolium Vahl. J. Ecol. 
51 : 727-732. 

Jaubert, H. J. Comte de, & Spach, E. 1842-43. J//ustrationes plantarum Orientalum 1. Paris. 

Jones, S. W. 1980. Morphology and major Taxonomy of Garcinia (Guttiferae). Unpublished thesis. 
Leicester University. 

Kapoor, B. M. 1972. Jn A. Love (Ed.), IOPB chromosome number reports, XXXV. Taxon 21 : 
161-166. 

Kawano, S. 1965. Anatomical studies on the androecia of some members of the Guttiferae- 
Moronoboideae. Bot. Mag. Tokyo 78 : 97-108. 

Keller, R. 1893. Hypericum. In A. Engler & K. Prantl, Die natiirlichen Pflanzenfamilien 3 (b) : 208- 
25: 

— 1925. Hypericum. In A. Engler & K. Prantl, Die nattirlichen Pflanzenfamilien 21 : 175-183. 2nd 
ed. 

Kemp, E. M. 1978. Tertiary climatic evolution and vegetational history in the southeast Indian Ocean 
region. Palaeogeogr., Palaeoclimat., Palaeoecol. 24 : 169-208. 

Kexel, H. 1896. Anatomie der Laubbldtter und Stengel der Hypericaceae und Cratoxyleae. Inaug.- 
Diss., Univ. of Erlangen. 

Killick, J. B. & Robson, N. K. B. 1976. Clusiaceae. Jn J. H. Ross (Ed.) Flora of Southern Africa 
22: 14-24. 

Kinzel, W. 1913. Frost und Licht als beeinflussende Krafte bei der Samenkeimung. Stuttgart. 

— 1920. Frost und Licht als beeinflussende Krafte bei der Samenkeimung. Nachtrag IT. Stuttgart. 

Knuth, P. 1908. Handbook of Flower Pollination 2 : 203-206. Oxford. [Transl. J. R. A. Davis.] 

Kobuski, C. E. 1950. Studies in the Theaceae, XIX. The genera Archytaea and Ploiarium. J. Arnold 
Arbor. 31 : 196-207. 

Krusheva, R. M. 1975. Jn A. Love (Ed.), IOPB chromosome number reports, L. Taxon 24 : 671-678. 

Kubitzki, K. 1969. Chemosystematische Betrachtungen zur Grossgliederung der Dicotylen. Taxon 
18 : 360-368. 

— 1977. The problem of rare and of frequent species: the monographer’s view. Jn G. T. Prance & T. 
S. Elias (Eds) Extinction is Forever : 331-336. New York. 

— 1978. In Maguire, B., The botany of the Guayana Highland—part X. Caraipa and Mahurea 
(Bonnetiaceae). Mem. N.Y. bot. Gdn 29 : 82-138. 

— Mesguita, A. A. L. & Gottlieb, O. R. 1978. Chemosystematic implications of xanthones in 
Bonnetia and Archytaea. Biochem. Syst. Ecol. 6 : 185-187. 

Kuntze, O. 1880. Miscellen tiber Hybriden und aus der Leipziger Flora (16). H. humifusum x per- 
foratum. Flora, Jena 63 : 293-294. 

Kyhos, D. W. 1967. Documented chromosome numbers of plants. Madrorio 19 : 134-136. 

Laane, M. M. 1969. Meiosis og kromosomstrukturell hybriditet i en del norske plantearter. Blyttia 
27 : 141-173. 

Lakhanpal, R. N. 1976. The antiquity of angiosperms. Aspects FI. Sci. 1 : 41-52. 

Leins, P. 1964. Das zentripetale und zentrifugale Androeceum. Ber. dt. bot. Ges. 77, Suppl. : 22-26. 

— 1971. Das Androeceum der Dikotylen. Ber. dt. bot. Ges. 84 : 191-193. 

Lessani, H. & Chariat-Panaki, S. 1979. Jn A. Love (Ed.) IOPB chromosome number reports, LXIV. 
Taxon 28 : 391-408. 

Lima, R. A. de, Gottlieb, O. R. & Lins Mesquita, A. A. 1972. Xanthones from Caraipa densiflora. 
Phytochem. 11 : 2307-2309. 

Linnaeus, C. 1753. Species plantarum. Stockholm. 


222 N. K. B. ROBSON 


Loon, J. C., van 1974. A cytological investigation of flowering plants from the Canary Islands. Acta bot. 
neerl. 23 : 113-124. 

— & de Jong, H. 1978. Jn A. Love (Ed.), I[OPB chromosome number reports, LIX. Taxon 27 : 53-61. 

— & Snelders, H. M. 1979. Jn A Love (Ed.), IOPB chromosome number reports LXIV. Taxon 28: 
Taxon 28 : 391-408. 

Love, A. & Kjellqvist, E. 1974. Cytotaxonomy of Spanish plants. III. Dicotyledons: Salicaceae— 
Rosaceae. Lagascalia 4 : 153-211. 

Maguire, B. 1970. On the flora of the Guayana Highland. Biotropica 2 : 85-100. 

— 1972. The botany of the Guayana Highland—part IX. Bonnetiaceae. Mem. N.Y. bot. Gdn 
23 : 131-165. 

Martin-Sans, M. E. 1922. Sur le polymorphism florale de l’ Hypericum humifusum L. Bull. Soc. Hist. 
Nat. Toulouse 50 : 214-227. 

Mathis, C. 1963. Etude chimio-taxonomique du genre Hypericum L. Thesis, Fac. Sci. Univ. 
Strasbourg. : 

— & Ourisson, G. 1963. Etude chimio-taxonomique du genre Hypericum |. Répartition de 
P’hypéricine. Phytochem. 2 : 157-171. 

—— —— 1964a. Etude chimio-taxonomique du genre Hypericum. II. Identification de constituants de 

diverses huilles essentielles d’ Hypericum. Phytochem. 3 : 115-131. 

—— —— 1964b. Etude chimio-taxonomique du genre Hypericum. Ill. Répartition des carbures saturés 
et des monoterpénes dans les huilles essentielles d’ Hypericum. Phytochem. 3 : 133-141. 

—— — 1964c. Etude taxonomique du genre Hypericum. IV. Répartition des sesquiterpénes des 
alcools monoterpeniques et des aldehydes saturés dans les huiles essentielles d’ Hypericum. Phyto- 
chem. 3 : 377-378. 

—— — 1964d. Etude taxonomique du genre Hypericum. V. Identification de quelques constituants 
non volatils d’ Hypericum perforatum L. Phytochem. 3 : 379. 

Matsuura, H. & Sato, T. 1935. Contributions to the idiogram study in phanerogamous plants. I. J. Fac. 
Sci. Hokkaido Univ. (Bot.)5 : 33-75. 

Mayr, E. 1968. Theory of biological classification. Nature 220 : 545-548. 

Mehra, P. N. & Sareen, T.S. 1969. Jn A. L6ve (Ed.), IOPB chromosome number reports, X XII. Taxon 
18 : 433-442. 

Melchior, H. 1930. Decaphalangium, eine neue Gattung der Guttiferen aus Peru. Notizbl. bot. Gart. 
Berlin 10 : 946-949. 

Melville, R. 1962. A new theory of the Angiosperm flower: I. The gynoecium. Kew Bull. 16 : 1-S0. 

1963. A new theory of the Angiosperm flower: II. The androecium. Kew Bull. 17 : 1-63. 

— 1966. Continental drift, Mezozoic continents and the migrations of Angiosperms. Nature 
211 : 116-120. 

—— 1969. Leaf venation patterns and the origin of the Angiosperms. Nature, Lond. 224 : 121-125. 

—— 1973. Continental drift and plant distribution. Jn D. H. Tarling & S. K. Runcorn (Eds) Jmpli- 
cations of Continental Drift to the Earth Sciences. I. London. 

— 1975. The distribution of Australian relict plants and its bearing on Angiosperm evolution. Bot. J. 
Linn. Soc. 71 : 67-88. 

— 1976. The terminology of leaf architecture. Taxon 25 : 549-561. 

— 1981. Vicarious plant distributions and the palaeogeography of the Pacific region. Jn G. Nelson & 
D. E. Rosen (Eds) Vicariance Palaeogeography. New York. 

Merrill, E. 1909. New or noteworthy Philippine plants, VII. Philipp. J. Sci. (Bot.) 4 : 247-330. 

Metcalfe, C. R. & Chalk, L. 1950. Anatomy of the Dicotyledons. Oxford. 

Meyer, F. K. 1978. Hypericum haplophylloides Hal. & Bald.—eine seltene Pflanze der westlichen 

Balkan-Halbinsel. Wiss. Z. Friedrich-Schiller- Univ. Jena, Math.- Nat. 27 (1): 67-78. 

Milne-Redhead, E. 1953. Hypericaceae. Jn W. B. Turrill & E. Milne-Redhead (Eds) Flora of Tropical 

East Africa. London. 

Moggi, G. & Pisacchi, A. 1967. Adumbratio florae aethiopicae. 14. Hypericaceae. Webbia 
22 : 233-289. 

Molly, E. 1875. Untersuchungen tiber die Bliithenentwickelung der Hypericaceen und Loasaceen mit 
besonderer Beriicksichtigung der verzweigten Staubgefdsse. Inaug. Diss. Univ. Bonn. 

Moore, D. M. 1973. In D. A. Webb & G. Halliday. The distribution, habitat and status of Hypericum 
canadense L. in Ireland. Watsonia 9 : 333-344. 

Mogquin-Tandon, A. 1826. Essaie sur les dédoublements ou multiplications d’organes dans les — 
végétaux. Montpellier. 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) DP! 


Morton, J. K. 1972. Phytogeography of the West African mountains. Jn D. H. Valentine (Ed.) 
Taxonomy, Phytogeography and Evolution : 221-236. London and New York. 

Miller, H. 1883. The Fertilisation of Flowers. London. [Trans. D’Arcy W. Thompson. ] 

Myers, O., Jr. 1963. Studies of Reproduction and Hybridization in Five Species of Hypericum Native 
to Eastern North America. Ph.D.Thesis, Cornell Univ. (See Diss.Abstr. 24 : 3515-3516). 

— 1964. Megasporogenesis, megagametophyte development, and endosperm development in 
Hypericum virginicum. Am. J. Bot. 51 : 664. 

Nelson, C. H. & Van Horn, G. S. 1976. A new simplified method for constructing Wagner networks 
and the cladistics of Pentachaeta (Compositae, Astereae). Brittonia 27 : 362-372. 

Nelson, G. 1976. Biogeography, the vicariance paradigm and continental drift. Syst. Zool. 24 : 
490-504. 

Nielsen, N. 1924. Chromosome numbers in the genus Hypericum. (A preliminary note.) Hereditas 
5 : 378-382. 

Nilsson, O. & Lassen, P. 1971. Chromosome numbers of vascular plants from Austria, Mallorca and 
Yugoslavia. Bot. Notiser 124 : 270-276. 

Noack, K. L. 1930. Untersuchungen tiber die Entstehung von Schecken und Chimaeren bei bunten 
Pflanzen. Rep. Proc. Intern. bot. Congr., Cambridge : 203-204. 

— 1931. Uber Hypericum-Kreuzungen. I. Die Panaschiire der Bastarde zwischen H. acutum Meench 
und H. montanum L. Z. indukt. Abstamm.-u. VerebLehre 59 : 77-101. 

— 1932a. Uber Hypericum-Kreuzungen. II. Beobachtungen an Hypericum-Artbastarden. Ber. dt. 
bot. Ges. 50 : 256-268. 

— 1932b. Uber Hypericum- Kreuzungen. III. Rassen- und Artkreuzungen mit einem buntblattrigen 
Hypericum acutum. Z. indukt. Abstamm.-u. VererbLehre 63 : 232-255. 

— 1934. Uber Hypericum-Kreuzungen. IV. Die Bastarde zwischen Hypericum acutum Ménch, 
montanum L., quadrangulum L., hirsutum L. und pulchrum L. Z. Bot. 28: 1-71. 

— 1937a. Uber Hypericum- Kreuzungen. V. Weitere Untersuchungen iiber die Buntblattrigkeit bei 
Artbastarden. Z. indukt. Abstamm.-u. VererbLehre 73 : 108-130. 

— 1937b. Weitere Untersuchungen iiber die Buntblattrigkeit bei Hypericum-Artbastarden. Z. 
indukt. Abstamm.-u. VererbLehre 73 : 373-375. 

— 1939. Uber Hypericum Kreuzungen. VI. Fortpflanzungsverhaltnisse und Bastarde von H. perfor- 
atum L. Z. indukt. Abstamm.-u. VererbLehre 76 : 569-601. 

— 1941. Geschlechtsverlust und Bastardierung beim Johanniskraut. Forsch. u. Fortschr. 17 : 13-15. 

Ohlendorf, O. 1907. Beitrdge zur Anatomie und Biologie der Fruchte und Samen einheimischer 
Wasser- und Sumpfpflanzen. \naug. Diss. Univ. Erlangen. 

Ornduff, R. 1975. Heterostyly and pollen flow in Hypericum aegypticum (Guttiferae). Bot. J. Linn. 
Soc. 71 : 51-57. 

Ortega, J. & Navarro, B. 1978. Estudios en la flora de Macaronesia: algunos numeros de cromosomas 
III. Botanica macaron. 3 : 73-80. 

Palm, B. 1922. Das Endosperm von Hypericum. Svensk bot. Tidskr. 16 : 60-68. 

Payer, J.-B. 1857. Traité d’organogénie compareée de la fleur. Paris. 

Perrier de la Bathie, H. 1951. 135. Hypericacées. Jn H. Humbert (Ed.) Flore de Madagascar et des 
Comores. Paris. 

Philipson, W. R. 1974. Ovular morphology and major classification of the dicotyledons. Bot. J. Linn. 
Soc. 68 : 89-108. 

— 1977. Ovular morphology and the classification of the dicotyledons. Plant Syst. Evol. Suppl. 
1: 123-140. 

Polya, L. 1950. Magyarorszagi novenyfajok kromoszomaszamai, II. Ann. Biol. Debrecen 1 : 46-56. 

Prakash, N. & Lau, Y. Y. 1976. Morphology of Ploiarium alternifolium and the taxonomic position of 
Ploiarium. Bot. Notiser 129 : 279-285. 

Prance, G. T. 1978. The origin and evolution of the Amazon flora. /ntersciencia 3 : 207-222. 

Pringle, J. S. 1976. Annotated chromosome counts for some plants of the dunes and pannes along the 
shores of the Upper Great Lakes. Mich. Botanist 15 : 157-163. 

Pringsheim, E. G. & Schwarz, W. 1933. Das Auftreten weissbunter (panaschierter) Pflanzen in der 
Natur. Flora, Jena 128 : 111-122. 

Rao, A. N. 1957. The embryology of Hypericum patulum Thunb. and H. mysorense Heyne. Phyto- 
morphology 7 : 36-45. 

Raven, P. H. 1979. Plate tectonics and southern hemisphere biogeography. In K. Larsen & L. B. Holm- 
Nielsen (Eds) Tropical Botany. London. 


224 N. K. B. ROBSON 


Raven, P. H. & Axelrod, D. I. 1974. Angiosperm biogeography and past continental movements. Ann. 
Mo. bot. Gdn 61 : 539-673. 

Rehder, A. 1910. Einige neue und kritische Gehdlze. Mitt. deut. dendr. Ges. 1910 : 248-254. 

— 1911. Pistillody of stamens in Hypericum nudiflorum. Bot. Gaz. 51 : 230-231. 

1940. Manual of Cultivated Trees and Shrubs : 640. 2nd ed. New York. 

Reid, C. & Reid, E. M. 1915. The Pliocene floras of the Dutch-Prussian border. Meded. Delfst. 6: 118, 
(i, 13). hess I/We). 

Reid, E. M. 1923. Nouvelles recherches sur les graines du Pliocéne inferieur du Pont-du-Gail (Cantal). 
Bull. Soc. géol. France IV, 23 : 305-355. 

Reitsma, T. 1969. Size modification of pollen grains under different treatments. Rev. Palaeobot. 
Palynol. 9 : 175-202. 

— 1970. Suggestions towards unification of descriptive terminology of angiosperm pollen grains. 
Rey. Palaeobot. Palynol. 10 : 39-60. 

Reynaud, C. 1973. Contribution a |’étude cytotaxinomique du genre Hypericum L. en Turquie. I. Bull. 
Soc. bot. France 120 : 201-216. 

— 1975. Contribution a |’étude cytotaxinomique de quelques Hypericum mediterranéens. Rey. Biol. 
Ecol. medit. 2 : 3-8. 

— 1980. In A Love (Ed.), IOPB chromosome number reports, LX VI. Taxon 29 : 163-169. 

Ridley, H. N. 1930. The Dispersal of Plants Throughout the World. Ashford, Kent. 

Robson, N. K. B. 1956. Studies in the Genus Hypericum L. Unpubl. thesis, University of Edinburgh. 

— 1957. Hypericum maculatum Crantz. Proc. botl Soc. Brit. Isles 2 : 237-238. 

— 1958a. Hypericum maculatum in Britain and Europe. Proc. botl Soc. Brit. Isles 3 : 99-100. 

— 1958b. The genus Hypericum in Africa south of the Sahara, Madagascar and the Mascarenes. 
Kew Bull. 12 : 433-446. 

— 1961. Guttiferae. Jn A. W. Exell & H. Wild (Eds) Flora Zambesiaca 1 : 378-404. 

— 1967a. Materials for a Flora of Turkey: XI. Notes on Turkish species of Hypericum. Notes R. bot. 
Gdn Edinb. 27 : 185-204. 

—— 1967b. Hypericum L. In P. H. Davis (Ed.) Flora of Turkey and the East Aegean Islands 
2 : 355-401. Edinburgh. 

— 1967c. Notes on Hypericum of S.W. Asia. Anz. Math.-Nat. Kl. Ost. Akad. Wiss. 104 (6): 1-8. 

— 1968a. 49. Guttiferae. In K. H. Rechinger (Ed.) Flora Iranica : 1-20. Graz. 

— 1968b. Guttiferae (Clusiaceae). In T. G. Tutin et al. (Eds) Flora Europaea 2: 261-269. 
Cambridge. 

— 1970. Shrubby Asiatic Hypericum species in cultivation. J. Roy. hort. Soc. 95 : 482-497. 

— 1972a. Evolutionary recall in Hypericum (Guttiferae) ? Trans. bot. Soc. Edinb. 41 : 365-383. 

— 1972b. Hypericum x desetangsii. Watsonia 9 : 203. 

— 1973a [‘1972’]. Notes on Malesian species of Hypericum (Guttiferae). Flora Malesian 
Praecursores, LII. Blumea 20 :251-274. 

— 1973b. Hypericum. In P. S. Green (Ed.) Plants: Wild and Cultivated : 79-82. London. 

— 1974. Hypericaceae. In C. G. G. J. van Steenis (Ed.) Flora Malesiana I, 8 : 1-29. 

— 1975. Hypericum L. In C. A. Stace (Ed.) Hybridization and the Flora of the British Isles : 164- 
167. London. 

— 1976. Guttiferae. Jn H.-L. Li et al. (Eds) Flora of Taiwan 2 : 620-646. 

— 1977a. Studies in the genus Hypericum L. (Guttiferae). I. Infrageneric classification. Bull. Br. Mus. 
nat. Hist. (Bot.) 5 : 291-355. 

— 1977b. Notes on some Nepalese and Indian Hypericum. J. Jap. Bot. 52 : 276-288. 

— 1978. Family 123A. Hypericaceae. Jn R. E. Woodson, R. W. Schery et al. (Eds) Flora of Panama. 
Ann. Mo. bot. Gdn 65 : 9-26. 

— 1979. Parallel evolution in African and Mascarene Hypericum. Kew Bull. 33 : 571-584. 

— 1980a. A new species of Hypericum from Angola. Bol. Soc. Brot. 11,53 : 113-121. 

— 1980b. Hypericum olympicum—wild and cultivated. The Plantsman 1 : 193-200. 

— 1980c. The Linnaean species of Ascyrum (Guttiferae). Taxon 29 : 267-274. 

— & Adams, W. P. 1968. Chromosome numbers in Hypericum and related genera. Brittonia 
20 : 95-106. 

— & Davis, P. H. 1980. An addition to the southern Turkish species of Hypericum sect. Arthro- 
phyllum. Notes roy. bot. Gdn Edinb. 38 : 103-104. 

Rodriguez-Jiménez, C. 1973. Recherches sur Hypericum L. section Brathys (Mutis ex L.f.) Choisy 
sous-section Spachium Keller (Guttiferae). Mem. Soc. Ci. nat. La Salle 33 : 5-151. 

Rosen, D. E. 1976. A vicariance model of Caribbean biogeography. Syst. Zool. 24 : 431-464. 


rq 


STUDIES IN THE GENUS HYPERICUM L. (GUTTIFERAE) 225 


Roth, L. 1953. Untersuchungen verschiedener Johanniskrautarten auf ihren Gehalt an Hypericin. Din 
Apoth.Ztg 36 : 653-656. 

Rury, P. M. & Dickison, W. C. 1977. Leaf venation patterns of the genus Hibbertia (Dilleniaceae). J. 
Arn. Arbor. 58 : 209-256. 

Sachs, F. G. J. von 1875. Textbook of Botany, Morphological and Physiological, Oxford. [Transl. A. 
W. Bennett & W. T. Dyer.] 

St Hilaire, A. F. C. P. 1824-28. Plantes usuelles des Brasiliens. Paris. 

Salisbury E. J. 1942. The Reproductive Capacity of Plants. London. 

— 1952. Downs and Dunes. London. 

— 1963. Fertile seed production and self-incompatibility of Hypericum calycinum in England. 
Watsonia 5 : 368-376. 

Sareen, T. S., Kant, S. & Pratap, R. 1974. Jn A. Love (Ed.), IOPB chromosome number reports, XLV. 
Taxon 23 : 619-624. 

Sattler, R. 1972. Organogenesis of Flowers. A Photographic Text-atlas. Toronto. 

— 1976. Centrifugal primordial inception in floral development. Ady. Pl. Morph. 1972 : 170-178. 

Saunders, E. R. 1936. On rhythmical development and radial organisation in the flower. J. Linn. Soc. 
(Bot.) 50 : 291-322. 

— 1937. The non-unique position in the genus Hypericum of H. peplidifolium A. Rich. Proc. Linn. 
Soc. 149 : 160-163. 

— 1939. The neglect of anatomical evidence in the current solutions of problems in systematic 
botany. New Phytol. 38 : 203-209. 

Schnarf, K. 1914. Beitrage zur Kentniss der Samenentwicklung einer europadischer Hypericum-Arten. 
Sber. Akad. Wiss. Wien, Math.-Nat. Kl. 1,123 : 159-188. 

Schofield, E. K. 1968. Petiole anatomy of the Guttiferae and related families. Mem. N.Y. bot. Gdn 
18: 1-SS. 

Schwarz, O. 1965. Die kritischen Hypericum-Arten der Mitteleuropadischen Flora. Drudea 5 : 59-66. 

Siersch, E. 1927. Anatomie und Microchemie der Hypericumdriisen. Planta 3 : 481-489. 

Simpson, B. B. 1975. Pleistocene changes in the flora of the high tropical Andes. Paleobiology 
1 : 273-294. 

Smith, A. C. 1973. Angiosperm evolution and the relationship of the floras of Africa and America. Jn 
B. J. Meggers, E. S. Ayensu & W. D. Duckworth (Eds) Tropical Forest Ecosystems in Africa and 
South America: a Comparative Review : 49-61. Washington. 

Smith, A. G., Briden, J. C. & Drewry, G. E. 1973. Phanerozoic world maps. Jn N. F. Hughes (Ed.) 
Organisms and Continents through Time : \-42. [Syst. Ass. Publ. No. 9] London. 

Souéges, E. C. R. 1925. Embryogenie des Hypericacées. Développement de |’°embryon chez H. perfor- 
atum. C. r. hebd. Séanc. Acad. Sci., Paris 180 : 949-951. 

— 1936. Embryogenie des Hypericacées. Développement de l’embryon chez |’Androsaemum 
officinale All. C. r. hebd. Séanc. Acad. Sci., Paris 202 : 679-681. 

Spach, E. 1836a. Hypericacearum monographiae fragmenta. Annis Sci. nat. (Bot.) Il, 5: 157-176, 
t. 4-5. 

— 1836b. Conspectus monographiae Hypericacearum. Annis Sci. nat. (Bot.) II, 5 : 349-369, t. 6. 

— 1836c. Histoire naturelle des végétaux. Phanérogames 5 : 335-464. 

Spirlet, M. 1967. Etude taxinomique des épidermes foliaires des Hypéricacées et des Guttiferacées du 
basin du fleuve Congo. Bull. I.F.A.N. A,9 : 5-91. 

Sporne, K. R. 1948. A note on a clearing technique of wide application. New Phytol. 47 : 290-291. 

Standley, P. C. 1940. Studies of American plants—XI. Guttiferae. Publ. Field Mus. nat. Hist., Chicago 
(Bot.) 22 : 159-161. 

Stauffer, H. S. 1963. Gestaltwandel bei Blutenstanden von Dictyledonen. Bot. Jb. 82 : 216-251. 

Stebbins, G. L. 1974. Flowering Plants. Evolution Above the Species Level. London. 

— & Hoogland, R. D. 1976. Species diversity. Ecology and evolution in a primitive Angiosperm 
genus: Hibbertia (Dilleniaceae). Plant Syst. Evol. 125 : 139-151. 

Steenis, C. G. G. J. van 1934. On the origin of the Malaysian mountain flora. Part 1. Facts and state- 
ment of the problem. Bull. Jard. bot. Buitenzorg, II, 13 : 1-417. 

— 1979. Plant-geography of east Malesia. Bot. J. Linn. Soc. 79 : 97-178. 

Stenar, H. 1938. Das Endosperm bei Hypericum acutum Moench. Bot. Notiser 1938 : 515-527. 

Stern, F. K. 1960. A Chalk Garden. London. 

Stout, G. H., Alden, R. A., Kraut, J. & High, D. F. 1962. Harunganin, a crystallographic deter- 
mination of an unknown structure. J. Am. Chem. Soc. 84 : 2653-2654. 


226 N. K. B. ROBSON 


Styer, C. H. & Stern, W. L. 1980. Comparative anatomy and systematics of woody Saxifragaceae. 
Deutzia. Bot. J. Linn. Soc. 79 : 291-319. 

Sugiura, T. 1940. A list of chromosome numbers in angiospermous plants, VI. Proc. imp. Acad. Tokyo 
16: 15-16. 

—— 1944. Studies on the chromosome numbers in higher plants. VI. Cytologia 13 : 352-359. 

Swamy, B. G. L. 1946. Endosperm in Hypericum mysorense. Ann. Bot. 10: 165-169. 

Synge, P. M. 1950. (Ed.) The proceedings of the Royal Horticultural Society. J. R. hort. Soc. 75 : c. 

Thellung, A. 1912. La flore adventice de Montpellier. Cherbourg. 

Thomas, J. L. 1970. Haploid and diploid pollen in Hypericum patulum. J. Arn. Arbor. 51 : 247-250. 

Thomson, G. M. 1922. The Naturalisation of Animals and Plants in New Zealand. Cambridge. 

Thorne, R. F. 1975. Angiosperm phylogeny and geography. Ann. Mo. bot. Gdn 62 : 362-367. 

Tieghem, P. van 1875. Recherches sur Ja structure du pistil et sur l’anatomie comparée de la fleur. 
Mem. Acad. Sci. France 21 : 1-261. 

Tillson, A. H. & Bamford, R. 1938. The floral anatomy of the Aurantioideae. Am. J. Bot. 25 : 780-793. 

Tucker, S. C. 1976. The role of ontogenetic evidence in floral morphology. Ady. Pl. Morph. 
1972 : 359-369. Meerut. 

Vautier, S. 1949. La vascularisation florale chez les Polygonacées. Candollea 12 : 219-343. 

Vestal, P. A. 1938. The significance of comparative anatomy in establishing the relationship of the 
Hypericaceae to the Guttiferae and their allies. Philipp. J. Sci. 64 : 199-256. 

Wagner, W. H. 1952a. The fern genus Diellia. Univ. Calif, Publs Bot. 26 : 1-212. 

— 1952b. Types of foliar dichotomy in living ferns. Am. J. Bot. 39 : 578-592. 

— 1969. The construction ofa classification. Nat. Acad. Sci. Publ. 1692 : 67-90. 

Webb, D. A. 1958. Hypericum canadense L. in Western Ireland. Watsonia 4: 140-144. 

— & Halliday, G. 1973. The distribution, habitat and status of Hypericum canadense L. in Ireland. 
Watsonia 9 : 333-344. 

Webb, D. H. 1980. A Biosystematic Study of Hypericum section Spachium in Eastern North America. 
Unpublished thesis. University of Tennessee, Knoxville. 

Wickens, G. E. 1976. Speculations on long distance dispersal and the flora of Jebel Marra (Sudan 
Republic). Kew Bull. 31 : 105-150. 

Wild, H. 1975. Phytogeography and the Gondwanaland position of Madagascar. Boissiera 
24: 107-117. 

Wilson, C. L. 1937. The phylogeny of the stamen. Am. J. Bot. 24 : 686-699. 

— 1953. The telome theory. Bot. Rev. 29 : 417-437. 

—— 1965. The floral anatomy of the Dilleniaceae, I. Hibbertia Andr. Phytomorphology 15 : 248-274. 

Wydler, H. 1851. Uber die symmetrische Verzweigungsweise dichotomischer Infloreszenzen. 24. 
Hypericineae. Flora, Jena 34 : 361. 

— 1859. Kleinere Beitrdge zur Kenntniss einheimischer Gewachse. Hypericineae. Flora, Jena 
42 : 364-367. 

— 1871. Kleinere Beitrage zur Kenntniss einheimischer Gewachse (Fortsetzung). Mitt. naturf. Ges. 
Bern 1871 : 52-54. 

— 1878. Zur Morphologie hauptsachlich der dichotomen Bliitenstande. J. wiss. Bot. 11 : 344-347. 

Zhukova, P. G. 1967. Kariologya nekotorikh rasteni pereselennikh v polyarno alpiyiskiy botanicheskii 
sad. In N. A. Avronin (Ed.) Pereselenie rastenii na Polynarnyi Serez IT : 139-149. 

Zimmermann, W. 1928. Histologische Studien am Vegetationspunkt von Hypericum uralum. Jb. wiss. 
Bot. 68 : 289-344. 

— 1959. Die Phylogenie der Pflanzen. 2nd ed. Stuttgart. 


British Museum (Natural History) 
1881-1981 


Centenary Publications 


Chance, change & challenge 


Two multi-author volumes from one of the foremost scientific institutions in the world. 


General Editor: P. H. Greenwood 


The Evolving Earth 


Editor: R. Cocks 


The Evolving Biosphere 
Editor: P. Forey 


In the first volume, The Evolving Earth, twenty scientists have been asked to summarise 
the present state of knowledge in their particular field, ranging from the origin of 
the Earth, through ocean sediments and soils to continental drift and palaeogeography. 


In the companion volume The Evolving Biosphere, museum scientists have chosen 
an evolutionary concept—speciation, coevolution, biogeography etc. and related 
this to the group of animals or plants in which they are specialising. Thus beetles 


and birds exemplify sympatric and alloptaric speciation, butterflies mimicry and 
certain fishes explosive evolution. 


In both volumes the text is supplemented by over one hundred specially-commissioned 
pieces of two-colour artwork. 


These two books will be invaluable to all sixth-form and undergraduate biology and 
geology students. 


The Evolving Earth: 276X219 mm, 280pp, 138 line illustrations, 42 halftones 
The Evolving Biosphere: 276219 mm, approx. 320pp, 133 line illustrations 
Publishing: Spring 1981 


Co-published by the British Museum (Natural History), London and Cambridge 
University Press, Cambridge. 


Titles to be published in Volume 8 


The Thelypteridaceae of Ceylon. By W. A. Sledge 


Studies in the genus Hypericum L. (Guttiferae) 
2. Characters of the genus. By N. K. B. Robson 


A revision of the lichen family Thelotremataceae in Sri Lanka. 
By Mason E. Hale, Jr. 


Vascular plant collections from the Tristan da Cunha group of 
islands. By Eric W. Groves 


Printed by Henry Ling Ltd, Dorchester 


Bulletin of the age 
British Museum (Natural History) 


A revision of the lichen family 
Thelotremataceae in Sri Lanka 


Mason E. Hale, Jr. 


| 
| 
_ Botany series Vol 8 No3 28 May 1981 


The Bulletin of the British Museum (Natural History), instituted in 1949, is issued in four — ay 
scientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology, and 
an Historical series. +g 


Papers in the Bulletin are primarily the results of research carried out on the unique and 


ever-growing collections of the Museum, both by the scientific staff of the Museum and by Bee 


works of reference that will remain indispensable for years to come. 


Parts are published at irregular intervals as they become ready, each is complete in itself, 

available separately, and individually priced. Volumes contain about 300 pages and several __ : 
volumes may appear within a calendar year. Subscriptions may be placed for one or more of — 
the series on either an Annual or Per Volume basis. Prices vary according to the contents of _ 
the individual parts. Orders and enquiries should be sent to: a. 


Publications Sales, 
British Museum (Natural History), 
Cromwell Road, 
London SW7 5BD, 
England. 


World List abbreviation: Bull. Br. Mus. nat. Hist. (Bot.) 


© Trustees of the British Museum (Natural History), 1981 


ISSN 0068-2292 Botany series i 
Vol 8 No 3 pp 227-332 

British Museum (Natural History) < 

Cromwell Road 


London SW7 5BD Issued 28 May 1981 


A revision of the lichen family Pee TIC HLALARIRO in 
Sri Lanka ; 


Mason E. Hale, Jr. 
Department of Botany, Smithsonian Institution, Washington, D.C. 20560, U.S.A. 


Contents 
SVMOPSIS. 0. <i. « « : : ee ee re eee D2 
Introduction . oo [ity eae hee a le ae Wk: 227 
Historical background . eo OG uae a at ee ere rena 228 
Morphology: structural characters Sor sely Oem sp ultalty F : ayer : 232 
SemuieciaGMaracicts -) oc 5) sy kl Uk ll 238 
Chemistry . ; eee ee oY re? 243 
Ecology and conservation “ipl eal al lla eda aa : ‘ ay <M. ; 244 
Do GIEE ENB yos SP pa ee ee 245 
SoeMemPecaiiicneee at) MRS BPP Fe 8 kk ED, 247 
Listofspecies .. Pee Creer eee Bare! ) wu! cal: so wel a 248 
PPROMPCHCEAMMEURGEiS Eh i ume ac cy sel 249 
eee see ee is RoE NY yes ee ee RSD 249 
MONO RTD RES  e  e 271 
ONE EOSIN ES ES a a es, 297 
SESE TERPS oe Re a ee ne 326 
a 326 
Index . 4 5 , : , é : ; : : : : : ; : : 328 

Synopsis 


This paper presents a revision of the lichen family Thelotremataceae in Sri Lanka based on the 
historical collections of Thwaites and Almquist and those of recent workers. The morphology, 
chemistry, phytogeography and ecology of the family are discussed and keys and descriptions for 110 
species presented. The following new species are described: Thelotrema dissultum Hale, T. 
imperfectum Hale, T. leprocarpoides Hale, T. nureliyum Hale, T. pidurutalagalum Hale, T. 
pulvereodiscum Hale, T. scabiomarginatum Hale, T. subpatens Hale, T. waasii Hale, Myriotrema 
albocinctum Hale, M. decorticatum Hale, M. fissurinum Hale, M. fluorescens Hale, M. frondosum 
Hale, M. mastarion Hale, M. multicavum Hale, M. nuwarense Hale, M. polytretum Hale, M. 
protoalbum Hale, M. thwaitesii Hale, Ocellularia albomaculata Hale, O. ascidioidea Hale, O. croceo- 
pora Hale, O. exuta Hale, O. kanneliyensis Hale, O. keralensis Patw. & C. Kulk. ex Hale, O. 
lankaensis Hale, O. meiospermoides Hale, O. melanotremata Hale, O. neocavata Hale, O. neopertu- 
sariiformis Hale, O. pluripora Hale, O. rassagala Hale, O. rhicnopora Hale, and O. sticticans Hale. 
One new combination, Ocellularia polillensis (Vainio) Hale, is made. 


Introduction 


One of the most remarkable yet little cited papers on tropical botany, ‘The lichens of 
Ceylon’, was published in 1870. In it the British lichenologist the Reverend W. A. Leighton 


Bull. Br. Mus. nat. Hist. (Bot.) 8 (3): 227-332 Issued 28 May 1981 


228 M. E. HALE 


listed nearly 200 species collected by G. K. Thwaites in Sri Lanka. A high percentage of 
these, some 44 species, were described as new. This pioneering effort in tropical lichenology 
was unfortunately not pursued further by Leighton, and he was soon eclipsed by the two 
giants of 19th century lichenology, W. Nylander and J. Miller (Argoviensis). Still, because 
Leighton antedates so much of their work, very few of his species have been relegated to 
synonymy, and they remain extremely important for tropical workers. The Thwaites’ 
specimens on which Leighton based his accounts are now in the herbarium of the British 
Museum (Natural History), London; they comprise an important part of the tropical lichen 
collections. 

Sri Lanka is an ideal place to conduct retrospective lichen research. With Thwaites’ 
collections as a base, we can determine the extent to which the lichen flora may have 
changed in the past 100 years through deforestation for agriculture and commercial logging. I 
visited Sri Lanka twice, in 1976 and in 1978, with the express purpose of collecting and 
studying the Thelotremataceae, an important tropical crustose family which figures 
prominently in Leighton’s flora. I was fortunate in having access to virgin dipterocarp forests 
being logged and succeeded in re-collecting most of the species reported by Leighton. At the 
same time I was able to add 76 species to the thelotreme flora to make a total of 110 species. 


Historical background 


Leighton’s Ceylon lichen flora 


The success of Leighton’s work must begin with Thwaites’ collections, for no herbarium of 
that day (and few even today) contained sufficient materials to prepare a revision of the 
lichens in any tropical country. G. H. K. Thwaites was appointed to the botanical post at 
Peradeniya, Sri Lanka, on 15 May 1849 by William Hooker. He served as director there from 
1857 to 1880 and died in Kandy in 1882. During his tenure he not only maintained the 
well-known botanical garden and herbarium but conducted extensive field excursions to 
most parts of Ceylon, very early complaining in his letters to Berkeley about the land leeches 
which still plague collectors in the rain forest. Almquist called him an enthusiastic naturalist 
(Nordenskidld, 1881). The accession books at the Royal Botanic Gardens in Kew are filled 
with entries of phanerogamic collections sent from 1850 to 1877. Many of these he had 
actually described himself while compiling a vascular flora of the island which eventually 
included 2832 species (Thwaites, 1864). 

Thwaites had broad interests in botany and started collecting fungi soon after he arrived. 
He wrote to Berkeley as early as 10 September 1850 that he was going ‘to send fungi and a 
few lichens this month’. They maintained active correspondence but lichens were only 
rarely mentioned again. Berkeley and Broome were eager to receive the fungi from this 
unknown tropical country and ultimately published on over 1000 species, describing 
hundreds as new (Berkeley & Broome, 1875) and encouraging Thwaites to continue 
collecting. 

Since there was no lichenologist at Kew Gardens, Berkeley and Broome asked Thwaites to 
send the lichens directly to Leighton in Shrewsbury without passing through Kew. Thwaites 
wrote in April 1867 that he was sending some lichens and learned from Berkeley in 
December that Leighton was indeed beginning to study the material. We know from 
Leighton’s letters to Nylander that three packages of lichens arrived between 25 November 
1867 and 19 November 1868, and by 23 February 1868 Berkeley was able to tell Thwaites 
that the lichens were being actively studied. Thwaites had stipulated that one set should go to 
Kew, one be retained by Leighton, and a third be given to Nylander. In fact, Leighton sent 
only selected fragments to Nylander and the remainder was ultimately deposited in Kew 
[now tranferred to BM; see Brenan & Ross, 1970] as part of his bequest. By January 1869 the 
finished paper, ‘The Lichens of Ceylon’, was read before the Linnean Society of London. In 


THELOTREMATACEAE IN SRI LANKA 229 


March 1872 Leighton told Berkeley that a few lichens remained to be described, but as far as 
I can determine he published nothing further on Sri Lanka. 

One cannot help but wonder how Leighton found any time to work on lichens at all. He 
was very active as pastor in the church in Shrewsbury and once reminded Nylander that he 
had three children to raise. His first botanical endeavour was the preparation of a 
cryptogamic flora of the county of Shropshire, followed by a phanerogamic flora published 
in 1841. His early scholarly works were on the angiocarpous lichens (1851) and the British 
Graphidaceae (1854), studies which obviously prepared him for a better understanding of the 
numerous crustose families in Sri Lanka. His chief work was actually the impressive The 
Lichen-flora of Great Britain, Ireland and the Channel Islands (1871), one of the classics of 
British botany and the forerunner of Crombie’s and later A. L. Smith’s ‘British Lichens’. 

A rather intimate picture of Leighton’s efforts to identify the Ceylon lichens, and the 
resources available to him, can be gleaned from his extensive correspondence with 
Nylander. From 1852 to 1874 he wrote about 120 letters to Nylander, often as frequently as 
one every two weeks. These are now part of the Nylander correspondence collection at the 
University of Helsinki library. 

Leighton wrote in English since he confessed poor French and no knowledge of German, 
the reason he gave for later cancelling the subscription to Flora, Jena which Nylander had 
arranged for him. On 3 December 1853, Leighton sent Nylander a complete catalogue of his 
lichen library, which included most of the published works on lichens of that day. Nylander 
very generously helped fill out these holdings for the next 15 years. Most of the letters are in 
fact concerned with details of books and articles he asked Nylander to find and methods of 
payment (banks then as now were reluctant to handle small drafts), as well as lists of species 
unrepresented in his collection that he needed to see. 

Leighton often enclosed small samples for comment or identification. Nylander helped 
identify the British Graphidaceae in this way, and many notes were exchanged on various 
species of Lecidea and other crustose genera, particularly between 1854 and 1865. Leighton 
also told Nylander of trips to Kew Gardens where he helped curate the Hooker Indian 
lichens. There must have been a small reference collection there since he looked for but 
could not find a specimen of Heterodea in 1866 and asked Nylander for one. He was as 
puzzled by this strange lichen as we are today. 

Leighton thus not only had the benefit of many specimens verified or authenticated by 
Nylander and any specimens on file at Kew but also received a complete set of Wright’s 
Cuban Lichens from Asa Gray in 1864, a set which is often cited in the ‘Lichen flora of 
Ceylon’. 

Nylander himself was unable to acquire this exsiccata and had to be satisfied with 
fragments from the set deposited in Paris. Leighton asked Nylander to add him to the list of 
subscribers for Lindig’s Colombian lichens (24 May 1864), which Nylander had exhaustively 
studied and published on in 1863. Leighton did not obtain a set but probably examined one 
in Kew since he also cites it in the Ceylon flora. He asked ‘Mr Krempelhuber’ for specimens 
of his ‘novelties’ but once again he did not receive any. There are no Leighton specimens at 
all in Krempelhuber’s herbarium (M), and Krempelhuber (1877) said that he saw none. 

I was able to locate 23 specimens of the Thelotremataceae of Ceylon which are now 
preserved in the Nylander herbarium in Helsinki. All are small fragments, no more than 
1-2 cm?, mounted on the same faintly ruled or blue letter paper Leighton used in his 
correspondence. He clearly wrote the ‘C. L.’ (Ceylon lichens) number assigned by Thwaites 
along with his own determination. Nylander often added ‘Thwaites’. Leighton told Nylander 
several times that he was only sending new species or otherwise interesting specimens so as 
not to burden him with too much work, and in fact samples of 10 species were not sent. In 
only one instance was a specimen sent without a name (C. L. 31, determined by Nylander as 
‘T. punctulatum Nyl.’). Once Nylander annotated a specimen (C. L. 99, Thelotrema 
sphinctrinellum) as ‘false determinato’, a notation which Leighton ignored, since this name is 
in the published list of Ceylon lichens, and in a similar case, C. L. 19 (Thelotrema albidi- 
forme Lgt. n. sp.) was redetermined correctly by Nylander as T. porinoides and later pub- 


230 M. E. HALE 


lished as such (Nylander, 1900). Leighton did not change his mind here either, although he 
thanked Nylander several times for his corrections of the Thwaites’ lichens. 

Two contemporary British lichenologists were often mentioned in the correspondence: 
Babington, much admired by Leighton, and Mudd, the author of an earlier lichen flora of 
Great Britain. His relationship with Mudd was evidently hostile, since Mudd had reviled 
him for adopting Nylander’s chemical tests (‘the absurdity of using chemistry in taxonomy’). 
Leighton almost gleefully and with obvious sarcasm passed Mudd’s comments on to 
Nylander and seemed to revel in his role as the proponent of chemical taxonomy in England. 
One should remember that it was Leighton who offered to translate Nylander’s now historic 
article on the use of hypochlorite of lime and caustic potash (Nylander, 1866). He was, in 
fact, the first lichenologist to adopt colour tests after Nylander. This was even noted by Th. 
M. Fries (1871: 58) when he critically discussed the use of KOH by Nylander to separate 
closely related Cladonias, adding, with obvious concern, that Leighton had willingly 
followed Nylander by using CaCl in a similar way. He questioned emphatically but futilely: 
‘Num ad species distinguendas idoneae sunt notae chemicae?’. Later Nylander (1891:100) 
bitterly criticized Fries for these comments. 

Leighton’s independence and confidence in handling the Ceylon lichen identifications 
may well explain the drop in correspondence with Nylander after 1869. Nylander was more 
accustomed to receiving unidentified lichen specimens which he would be free to describe as 
new, being without question the leading lichenologist of the day. Leighton, however, did not 
even write to him about the publication of the Ceylon flora, even though he told Nylander on 
12 April 1870 that he was working steadily on the ‘Lichen-flora of Great Britain’ and sent 
him a copy on 22 September 1871. In one of the last letters to Nylander (26 February 
1874), Leighton replied ‘to your not very courteous letter’, that he was ‘not quite so ignorant 
of “elementary botany” as not to know that zoospheres and spermatia are different 
organisms’, obviously reacting to criticism from Nylander. As the final insult, Nylander 
transmitted his own paper in opposition to Schwendenerism to Crombie (1891), even though 
Leighton had earlier offered to translate it and get it inserted in Grevillea (‘Your transfer of 
your paper to Mr Crombie only excites a smile’). 

Leighton finally gave up botany in 1879 because of failing eyesight and turned all of his 
collections over to Kew. He died 25 February 1888 (Phillips, 1889). 

Little reference has been made to Leighton’s Ceylon lichens since the 1870 publication. 
Krempelhuber (1877) was the first to cite them, without actually seeing the specimens, when 
he included five species described by Leighton in a monograph of Ascidium (A. granulosum, 
A. pachystomum, A. pertusariiforme, A. punctulatum, and A. thelotremoides). Nylander, as I 
mentioned above, used the names Asteristion erumpens and Thelotrema albidiforme (as a 
synonym of T. porinoides) in his list of the Almquists’ collections. Although Vainio (1921) 
identified 30 species of thelotremes from the Philippines, 24 of them new, he referred to only 
one Leighton species, T. pertusariiforme, in a discussion of T. curranii Vain. He had seen the 
Ceylon specimen in Nylander’s herbarium. As a matter of fact, I have identified only one 
Leighton species, Ocellularia chonestoma, from the Philippines. 


The Thwaites’ collections 


Thwaites shipped his well-prepared and ample collections to Leighton glued on white sheets. 
For reasons we cannot even guess now, Leighton remounted virtually all of these on a heavy 
blue letter paper, transferring the label data in his own hand and discarding the original 
labels. In contrast, Berkeley and Broome kept the fungi on their original mounts. The 
number of unicate thelotreme specimens sent added up to at least 49, and there is reason to 
believe that more were sent but not accounted for. It is also unfortunate that Leighton 
ignored Thwaites’ collection numbers in his final publication and simply numbered the 
species serially. One can be thankful that there is only rarely any confusion as to the 
collection used in the publication. 

The localities for the Thwaites’ lichens cannot be determined accurately except when he 


THELOTREMATACEAE IN SRI LANKA 231 


specifically refers to Peradeniya or Hakgala (‘Habgalla’). The most frequent designations, 
‘Central Province’ and ‘South of island’ are of no help. Notwithstanding, he collected many 
species, such as Ocellularia chonestoma, from the leech-infested low-elevation rain forests, 
along with obvious high-elevation species in the Horton Plains area, such as O. fissa. 

After Thwaites received Leighton’s published article, he transferred the names to sets of 
duplicate specimens preserved in Peradeniya along with a large series of duplicates for some 
of the numbers. Unfortunately, he further obscured the accuracy of the locality data by citing 
each species according to Leighton’s published species numbers (1-196), omitting all other 
information on the original C. L. (Ceylon Lichens) numbers or localities, even though he 
prepared a list of the corresponding numbers for his own purposes. I discovered a notebook 
in the Peradeniya herbarium with a partial list of the institutions where sets were to be sent, 
as follows: a: [blank]; b: British Museum 1875 [BM specimens were subsequently stamped 
‘Ceylon, G. H. K. Thwaites, Rec’d. 1875’]; c: [blank]; d: Geneva Prof. deCandolle for J. 
Miller 1876; e, f, and g: [blank]; h: Florence Roy. Acad. Sr. Parlatore Nov. 1875;i: J. K. 1.1 
have now been able to determine that sets were finally sent to British Museum (21 
specimens), Uppsala (UPS) (23 specimens), Stockholm (S) (20 specimens sent in 1879), 
Geneva (G) (11 specimens), Vienna (W) (5 specimens), and Paris (PC) (10 specimens) 
(numbers for G. PC, and W are counts of only type duplicates). These institutions would 
account for most of the blanks in the notebook. As I will show in the species lists under 
specimens examined, these duplicates are sometimes mixtures and the determinations not 
always trustworthy. 

The collections at Peradeniya are extremely important since 175 of the total 196 Leighton 
species, including most of the types, are represented there. It is unusual, even today, for 
tropical herbaria to have duplicates of lichens of their own country collected and determined 
by lichenologists from America and Europe. The only species missing are numbers 3, 8, 11, 
33, 34, 38, 42, 53, 57, 60, 62, 76, 101, 109, 125, 127-129, 173, and 191, as enumerated in 
Leighton’s publication. 

Thwaites apparently kept no field books for the C. L. number series which was reserved for 
lichens. Taking the Thelotremataceae, however, we find that the first C. L. collection 
number was 5 and the last 284. The last number in BM, however, is 175, and I am not sure 
how or when the C. L. numbers between 175 and 284, now filed on the same sheets in 
Peradeniya with numbers 5-175, were identified. In general, the correspondence between 
the PDA and BM collections, where both are accounted for, is surprisingly good. There are 
very few cases of erroneous mixtures such as Thelotrema exanthismocarpum (C. L. 97), 
which is represented in PDA by a misidentified ‘Ocellularia’ stictidea (C. L. 200) 
(= Conotrema). 


The Almquist collections 


In summarizing the historical development of research on the lichens of Sri Lanka, one 
cannot overlook a small but important collection of thelotremes made by Almquist, the 
famous Swedish explorer. He visited Peradeniya in December 1879. Thwaites, extremely 
pleased to have a distinguished visitor in this isolated post, took him on trips to the 
Peradeniya area, Point de Galle, a coastal area in southernmost Sri Lanka, and Piduru- 
talagala (Pedrotallegalle), the highest point on the island. I was able to climb this mountain 
in 1978 on the same well-worn trail used by Almquist and recollect some of the species he 
found. 

The Almquist collections, including those from Japan and other countries visited during 
the Vega expedition, were sent to Nylander for identification. This material formed the basis 
for Nylander’s final scientific work, ‘Lichenes Ceylonenses’, published posthumously in 
1900. The original specimens are now preserved in Stockholm with fragments in Nylander’s 
herbarium at Helsinki. In the article Nylander lists only two species described by Leighton, 
Asteristion erumpens (= Thelotrema platycarpum) and Thelotrema albidiforme (redeter- 
mined correctly as T. porinoides), both of which Leighton had sent him for comment. Seven 


232 M. E. HALE 


other species were collected by Almquist. When later preparing his own article on Ceylon 
lichens based on Almqujst’s collections, Nylander (1900: 26) admitted that his list was much 
poorer than Leighton’s and at the same time complained indirectly that he was 
not able to lay his hands on Leighton’s specimens, unfortunate in his mind since he had 
doubts about the identifications. This is a curious reprimand since Leighton had sent him 23 
Thwaites specimens of Thelotremataceae for comment in 1868-69. Representatives of the 
other families were also sent, although 1n lesser numbers. 


Recent collections 


Following Thwaites by almost 70 years, A. H. G. Alston spent the years 1926-31 in Sri 
Lanka, collecting widely and adding plant specimens, including a few lichens, to the 
Peradeniya herbarium (Crabbe, 1960). He wrote the Kandy flora (with a listing Thelotrema 
lepadizum) (Alston, 1938), a supplement to Trimen’s Handbook, and unpublished 
taxonomic treatments of Ceylon bryophytes, algae, and lichens. The treatment for lichens, 
now on file in the cryptogamic library of the British Museum, 1s essentially a compilation of 
Leighton’s lichen flora. 

Other recent collecting has been done incidentally under the Smithsonian Institutions’ 
Flora of Ceylon Project from 1970-1976, mostly by Louis Wheeler in drier lowland areas. 
His specimens are preserved in Peradeniya and the Smithsonian Institution. 

Among professional lichenologists, I know of a brief visit in 1975 by Rolf Santesson and 
Roland Moberg, who together collected about 10 specimens of Thelotremataceae, mostly in 
the Horton Plains area. These are in Uppsala with some duplicates in the Smithsonian (US). 
My own excursions are given below under ‘Collecting localities’ (p. 247). 

Other lichenologists and bryologists have probably collected the Thelotremataceae on Sri 
Lanka but I have not attempted to examine their collections. 


Morphology: structural characters 


We are still not sure what value should be attached to structural characters in the taxonomy 
of crustose lichens. Foliose lichens have great diversity not only in cortical structure, which 
lichenologists are only now taking into serious consideration, but also in vegetative struc- 
tures such as rhizines and diaspores such as isidia and soredia, which have long been 
accepted. Magnusson (1939: 10), the acknowledged master of crustose lichens, placed little 
value on cortical structure since ‘the real arrangement of the hyphae is difficult to observe in 
hand-cut sections and certainly is of little practical importance, irrespective of the trust- 
worthiness of this character’. It can indeed be difficult to interpret cortical organization in 
the Thelotremataceae and other crustose lichens with a light microscope (see Fig. 1) because 
the tissues are much denser than corresponding ones in macrolichens, but at the same time 
no character should be rejected simply because of difficulties in observing it. 

I have approached this problem by studying and comparing cross sections with both the 
light microscope and the scanning electron microscope, using the same specimen face when 
sectioning. Specimen preparation for the scanning microscope follows Hale (1973a: 2). The 
results are not claimed to be final but will hopefully point out interesting avenues for future 
research in the family. 


Hypophloeodal species 


According to various authors, hypophloeodal species lack a clearcut cortex, the algae being 
superficial or dispersed among the outer periderm layers. The medulla is mainly located 
among periderm cells. The scanning electron microscope supports this description in large 
part (Figs 2a, b; Hale, 1974a: 6, Fig 4), and the following species in Sri Lanka can be 
called hypophloeodal: Ocellularia leuacomelaena, O. meiosperma, O. meiospermoides, O. 
melanotremata, O. tenuis, Thelotrema alborosellum, T. coccineum, T. colobicum, T. 


THELOTREMATACEAE IN SRI LANKA 233 


Fig. 1 Thallus in the Thelotremataceae viewed with a light microscope. (a) Cross section of 
Myriotrema multicavum (Hale 46 150, x 500) showing the columnar medulla and crystals. (b) 
Exfoliating cortex of M. terebrans (Hale 47 202, x 500) with crystal inclusions. 


dilatatum, T. lacteum, T. lepadinum, T. leprocarpoides, T. monosporum, T. nostalgicum, 
T. novae-zelandiae, T. patens, T. phlyctidioides, T. porinoides, T. subpatens, T. waasii, and 
T. weberi. No species of Myriotrema are hypophloeodal although M. decorticatum has 
extensive decorticated areas. 

The algal layer in these species may actually persist intact among a thin, non-corticated 
layer of ‘medullary’ hyphae that covers the plant body as a wispy white colouration. In 
addition, the medullary hyphae are not only intercellular in the periderm, but also 
intracellular (Hale, 1974a: 6, Fig. 4d; also Fig. 2c). One soon realizes that hypophloeodal 
species may have no more medullary hyphae growing below the periderm than do the 
epiphloeodal (corticate) species. The term non-corticate approximates the true anatomical 
condition of these lichens most accurately, although a few distinctly thalloid species (as 
Myriotrema glaucescens) to be mentioned below may also lack a cortex. 

Obviously much work remains to be done before a full understanding of hypophloeodal 
species is achieved, not only in terms of morphological development but also correlations 
with other characters. For example, non-corticate species have an unusually high frequency 
of stictic acid or no substances. Furthermore, they lack completely psoromic acid and other 
substances often found in the corticate (epiphloeodal) species. These few observations 
suggest that lack of a cortex is a primitive character. 


Epiphloeodal species 
Cortex 


Epiphloeodal species have a distinct cortex, algal layer, and medulla (Fig. 2d). The scanning 
electron microscope shows more variation than one would expect from the usual hand 
drawings of the species (see Redinger, 1936, or Salisbury, 1978). Basically, the cortex consists 
ofa highly agglutinated, continuous layer 2-5 cells (8-15 um) thick, the individual cells lying 
more or less randomly in a horizontal plane, their walls heavily gelatinized. Salisbury’s term 
‘cartilaginous’ seems quite appropriate (see Figs |b, 6a, b). 

On closer inspection, one can detect subtle differences in the degree of gelatinization and 
density of packing, obviously characters which may intergrade and require subjective 
judgement. For example, the following species in Sri Lanka have rather loosely packed, 
individually recognizable cells in the cortex: Myriotrema anamalaiense, M. andamanicum, 
M. mastarion, M. porinaceum, M. thwaitesii, Ocellularia ascidioidea, O. chonestoma (Fig. 
3c), O. dolichotata, O. eumorpha, O. neopertusariiformis, O. nylanderiana, O. orthomastia, 


234 M. E. HALE 


Fig. 2 Thallus structure in the Thelotremataceae viewed with the scanning electron microscope. (a) 
Low power view of cross section of a hypophloeodal species (Thelotrema waasii, Hale 51 175). 
(b) Cross section of a hypophloeodal species showing patches of medullary hyphae (arrow) on 
periderm surface (T. dilatatum, Hale 46 218). (c) Enlargement of periderm cells to show intra- 
cellular hyphae (arrow) (Ocellularia kanneliyensis, Hale 46 213). (d) Low power view of cross 
section of an epiphloeodal species (O. perforata, Hale 47 010) showing cortex (C), algal layer (A), 
medulla with crystals (M), and periderm (P). 


O. pertusariiformis, O. rhicnopora, O. triglyphica, Thelotrema imperfectum (Fig. 3a), and T. 
nureliyum. The surface of some of these species may appear to be irregularly pored (Fig. 
3¢)} 

A second group of species has more heavily gelatinized, compressed cells in the cortex 
with only the small lumina suggesting outlines of hyphal organization: Myriotrema 
elachistoteron, M. granulosum, M. microporum, M. nuwarense, Ocellularia aurata, O. 
diacida, O. exuta, O. kanneliyensis, O. keralensis, O. lirelliformis, O. monosporoides, O. 
neocavata, O. sticticans, O. thelotremoides, Thelotrema aggregatum, T. astroideum, T. 
dissultum, T. kamatii, T. leprieurii, T. platycarpoides, T. platycarpum, T. pseudo- 
exanthismocarpum, and T. pulvereodiscum. The surface of these species is usually 
featureless and non-pored under the scanning electron microscope. 

A third group of species, comprising the majority of those in the family, is corticate but 
differs from the species mentioned above in having aculeate hyphae on the surface. I first 


THELOTREMATACEAE IN SRI LANKA 235 


Fig. 3 Cortical structure of the Thelotremataceae viewed with a scanning electron microscope. (a) 
Loosely organized cellular cortex of Thelotrema imperfectum (Hale 46 338). (b) Dense cellular 
cortex of Myriotrema album (Hale 4 171). (c) Loosely organized, irregularly pored cortex of 


Ocellularia chonestoma (Hale 46 344). (d) Dense cellular cortex of O. crassa with large crystal 
inclusions (Hale 51 210). 


discovered these peculiar hyphae in some West Indian species (Hale, 1974a: 5, Fig. 3), using 
the scanning electron microscope, and can now add more details from the larger material in 
Sri Lanka. Aculeate hyphae are unicellular, non-gelatinized, and erect and seem to fade into 
the main polysaccharide cortical surface (Fig. 4a). They are far too delicate and translucent 
to be seen with a light microscope and usually collapse after vacuum treatment needed in 
electron microscope preparation. 

There are three major groups of species with aculeate hyphae: (1) species with a loosely 
organized cortex (Ocellularia chonestoma, O. dolichotata, and O. pertusariiformis already 
mentioned above); (2) species with a dense, continuous cortex (Myriotrema albocinctum, M. 
eminens, M. fissurinum, M. hartii, M. olivaceum, M. protoalbum, Ocellularia crassa (Fig. 
3d), O. fissa, O. perforata, O. punctulata, and Thelotrema scabiomarginatum); and (3) 
species with a dense cortex which splits into sheets or layers and may exfoliate at the surface 
(Myriotrema album (Fig. 3b), M. cinereoglaucescens, M. clandestinum, M. costaricense, M. 
desquamans, M. fluorescens, M. frondosum, M. glaucophaenum, M. masonhalei, M. 
microstomum, M. minutum, M. polytretum, M. santessonti, M. subconforme, M. terebrans, 
M. terebratulum, Ocellularia albomaculata, O. croceopora, O. emersa, O. lankaensis, O. 


a a - all 


2.22KX 19.5KU 5.UM @29° #2 #027 © 
— : = ene oe ” “ e 


= ys 


Seana rin 
7 


Fig. 4 Cortical structure of the Thelotremataceae viewed with the scanning electron microscope. 
(a) Splitting cortex and aculeate hyphae of Myriotrema glaucophaenum (Hale 47 062). (b) 
Exfoliating sheet of Ocellularia emersa (Hale 51 068). (c) Internally splitting cortex of 
Myriotrema fluorescens (Hale 47 152). (d) Internally splitting cortex of Ocellularia croceopora 
(Hale 46 186) showing incorporation of sheets into the medulla. 


marivelensis, O. massalongoi, O. papillata, O. pluripora, O. polillensis, O. rassagala, O. 
subsimilis, Thelotrema magnificum, T. piluliferum, and T. platysporum). 

Exfoliation of the cortex was probably first observed by Miiller Argoviensis when he 
described Anthracothecium (Myriotrema) desquamans as having ‘plaguliformi desquamans’. 
Using a light microscope one can see the cortex splitting into thin sheets which may curl up 
and peel off the surface (Fig. 1b). The cortex may also split internally, giving rise in extreme 
cases to alternating, stratified layers of cortex and medulla, as in Ocellularia croceopora (Fig. 
4d). These layers may be only one cell layer thick or 20 um or more, regenerating from below 
as Sheets subdivide or peel off. The constant occurrence of aculeate hyphae in species with an 
exfoliating cortex suggests that these hyphae initiate or promote the splitting process in some 
way since they can be seen on the surface of newly exfoliating sheets (Fig. 4b). The same 
phenomenon occurs in the Graphidaceae (Wirth & Hale, 1978) and will surely be discovered 
in other crustose groups. 

There is a high degree of correlation between cortical structure and chemistry, although I 
have not tested this statistically. Species with hypoprotecetraric acid or no substances appear 


THELOTREMATACEAE IN SRI LANKA 237 


. { - ~~ é 
I be 


125% 19.5K0 ope ae, sd18 


> 


=e ae nk 85 eh : 
Yew ve S SS 
me - yeas : 


5 4 Pan f 


Fig. 5 Medullary structure of the Thelotremataceae viewed with the scanning electron microscope. 
(a) Upper medullary area of Myriotrema glaucescens (Hale 50 334) showing massed suberect 
hyphae at the surface. (b) Columnar cortex of M. microporum (Hale 46 271). (c) Crystal-filled 
medulla of M. multicavum (Hale 46 150) (compare with Fig. la). (d) Crystals of psoromic acid 
on medullary hyphae of Ocellularia lankaensis (Hale 51 008). 


to have a more loosely organized cortex and rarer occurrence of aculeate hyphae than 
expected. The correlation between aculeate hyphae, cortical splitting, and chemistry is more 
striking. All 23 species in Sri lanka with psoromic acid, excepting Myriotrema 
andamanicum and M. mastarion, have aculeate hyphae and a high frequency of cortical 
splitting. More exhaustive examination of all species in the family will undoubtedly reveal 
additional correlations. 


Algal layer 


The mycobiont in the Thelotremataceae is Trentepohlia. Within the thallus this alga forms 
unicellular spheres evenly distributed in a layer 24 cells thick (10-30 um). In many species, 
the layer is interrupted by intrusions of oxalate crystals. When the medulla is missing, the 
algal layer lies on and may be partly dispersed in the periderm. 


Medulla 


The medulla of epiphloeodal species varies considerably in thickness. Even in the same 


238 M. E. HALE 


plant it may be virtually absent—the algae resting directly on the periderm—or as much as 
60-100 um thick. The medulla is modified by the bark substratum in many ways. A uniform 
layer of medulla forms over a hard periderm, but if the periderm disintegrates easily the 
medulla may not only penetrate deeply in periderm fissures but even incorporate layers of 
bark cells. 

Many species of the Thelotremataceae produce oxalate crystals (Figs la, Sc), which are 
so abundant in some (e.g. Myriotrema multicavum) (Fig. 5c) that cohesion of the hyphae is 
broken and the thallus abrades and crumbles easily. The origin of these crystals is uncertain 
but from my observations they appear to be produced by the medullary hyphae and are 
species specific. They are probably a waste product, stimulated by the very close contact 
between the medulla and periderm. They may even have a conceivable function in 
magnifying the light rays which pass through the cortex and become dispersed to the algae 
closely surrounding the crystals. Species containing psoromic acid may have additional 
conspicuous masses of acid crystals on the hyphae (Fig. 5d). 

There is, finally, an unusual, highly developed medulla in a small group of species, all in 
the genus Myriotrema. It consists of vertically oriented, cellular blocks of tissue 
200-1000 wm thick (see Hale, 1974a: 3, Fig. 1d) with algae oriented between the vertical 
arrays. One such species in Sri Lanka, Myriotrema microporum, has a well developed cortex 
(Fig. 5b), but the others in this group, M. compunctum, M. glaucescens (Fig. 5a), M. 
multicavum, and M. wightii all lack a cortex. The surface consists of knobby, packed 
medullary hyphae sometimes covered with an irregular, pored polysaccharide sheet. I do not 
believe, however, this can be compared with the epicortex of foliose lichens. 

In conclusion, I feel that structural characters in crustose lichens, as exemplified by the 
Thelotremataceae, are sufficiently distinctive and constant to play some role in the 
taxonomy. However, we have barely begun to explore the nature and range of variation of 
these structures with the new analytical tools available; future workers will be in a better 
position to evaluate them. 


Apothecial characters 
The exciple 


Over 100 years ago Tuckerman (1864: 270) stated that ‘it will not be easy to found any 
subdivision of (7helotrema) on the presence or absence of the interior exciple’. There has, 
indeed, been much controversy and uncertainty about the usefulness of apothecial characters 
in the family even up to the present time. The only developmental study is one by 
Letrouit-Galinou (1966) for Thelotrema lepadinum, and many of our conclusions on the 
other 450 species in the family must be based on far less satisfactory evidence. 

To summarize what we do know about apothecial characters, one of the most 
characteristic features of the family is the periphysoid-bearing exciple, constant for the genus 
Thelotrema. This exciple arises from the base of the hymenium and is covered with 
periphysoids on the inner surface (Salisbury, 1972a). Periphysoids are short paraphyses-like 
hyphae oriented at right angles to the paraphyses of the hymenium (Fig. 6c), similar to the 
better known periphyses of pyrenocarpous lichens (Henssen & Jahns, 1974). Redinger (1936) 
overlooked them completely in his otherwise careful studies of the family in Brazil. They 
were apparently first recognized by Zahlbruckner (1941: 16) in Thelotrema periphysatum 
Zahlbr. (= T. lepadinum) and soon thereafter by Magnusson (Magnusson & Zahlbruckner, 
1944: 53) in Thelotrema gibbosum Magn. (= T. piluliferum Tuck.). 

Another typically thelotremoid character is the carbonized columella. While I know of no 
developmental studies on this structure, Redinger (1936: 7) found that the ascogonial initials 
(‘die Primordien der Apothecien’) of columellate species originate in the periderm at a depth 
of 3-5 cells. He surmised that the columella functions as a ‘Rammbock’, which breaches this 
layer of bark cells as the apothecia mature and reach the surface. Although this simplistic 
explanation may be close to the truth, a number of Myriotrema species (e.g. M. 


THELOTREMATACEAE IN SRI LANKA 239 


Fig. 6 Apothecial structure in the Thelotremataceae viewed with a light microscope. (a) Immersed 
apothecium of Myriotrema album (Hale 46 171). (b) Immersed apothecium of M. minutum 
(Hale 50 315). (c) Exciple with periphysoids (arrow) in Thelotrema nureliyum (Hale 50 292). (d) 
Thalline rim (T) and partially fused exciple (E) of 7. subpatens (Hale 46 208) (all about x 500). 


andamanicum and M. mastarion) have deeply embedded apothecia capable of reaching the 
surface without a columella. In any event columellate species always have a fused, 
carbonized, exciple lacking periphysoids. All species of Ocellularia, as circumscribed in this 
revision, have a carbonized exciple and a high proportion of them have a columella. A few 
species develop a broad, divided, or reticulate columella, often with an open disc. Vainio 
(1921: 184) recognized these in separate genus, Rhabdodiscus (= Stegobolus Mont.). 

A third major group of species, all in Myriotrema, have a fused, noncarbonized exciple 
without periphysoids (Figs 6a, b). The exciple is colourless or pale yellowish to reddish 
brown but may be poorly developed to lacking laterally. 

There is considerable variation in the development of the thalline wall. It is perhaps most 
conspicuous in the genus Thelotrema, where many species have a nearly erect rim and 
gaping pore (see Figs 7e, f). Other species of Thelotrema have an erect to strongly recurved 
and split wall (see Figs 9h, 10b), so unique that Leighton described a new genus Asteristion 
on the basis of one of them. I have called these apothecia chroodiscoid (Hale, 1978a: 2), 
deriving the term from the well-known genus name Chroodiscus (= Thelotrema). 

Ocellularia has a well-developed thalline wall described as incurved, resulting in a 
constricted, often rimmed pore. Myriotrema, on the other hand, generally has a weakly 


240 M. E. HALE 


developed wall or none at all, since many of the species have apothecia completely 
immersed in the medulla. 


Spore characters 


The family Thelotremataceae is characterized by four basic spore types differentiated by 
colour (colourless or brown) and septation (transversely septate or muriform). There is 
surprisingly little intergradation between them. In most cases, abnormal darkening of 
otherwise colourless spores seems to result from senescence. Ocellularia patwardhanii Hale 
(Hale, 19786: 379), O. rhodostroma (Mont.) Zahlbr., and O. subcavata (Nyl.) Zahlbr. (Hale, 
1974a: 28) may all have some brown spores at an evidently moribund stage. A few species 
such as Ocellularia monosporoides (Ny1.) Hale have predominantly colourless spores which 
turn brownish tardily at maturity. As for septation, a very few species with transversely 
septate spores may develop a few longitudinal septa in the end cells, as in Thelotrema 
pidurutalagalum. Abnormal spore pigmentation and septation are also known in the 
Graphidaceae (Wirth & Hale, 1978: 2). 

Some lichenologists have placed considerable emphasis on the shape of cell lumina in the 
family. For example, most species with 4-celled spores have thickened walls, giving a 
lenticular appearance. Densely muriform spores, as in Thelotrema monosporum, have more 
thin-walled, cubical cells. Obviously a definitive determination of shape of lumina will 
require much more detailed investigation. 


Generic classification 


Generic and subgeneric divisions of the Thelotremataceae have varied widely with different 
authors. Acharius proposed the first genus, Thelotrema, which included only 15 species in 
his lifetime. He was followed by Fée in 1824, who examined a wider range of specimens and 
had a greater appreciation of microscopic characters, recognizing three genera, Thelotrema 
and two others described as new, Ascidium and Myriotrema. These names were no sooner 
published than G. F. W. Meyer in 1825 added three more, Ocellularia (later conserved 
against Ascidium), Antrocarpum (based on Lichen lepadinus Ach. and therefore a synonym 
of Thelotrema), and Porophora (a superfluous name for Ascidium). 

Additional names were proposed by Massalongo, Montagne, Hampe, and others in the 
19th century, as will be seen in the lists of generic synonymy in the species list below, but few 
were ever adopted in practice. Nylander, the most active worker in the Thelotremataceae, 
began his studies by recognizing Ascidium and Thelotrema but soon reverted to one genus, 
Thelotrema, and during his long career from 1858 to 1900 finally ignored all other names, 
making no contributions at all to generic concepts in the family. He subdivided the species 
into four groups differing principally in spore colour, size, and septation (Nylander, 1862). 
His countryman, Vainio, also classified all species in Thelotrema, excepting the reticulate- 
columellate species, segregated as Rhabdodiscus. 

Miiller Argoviensis, Nylander’s contemporary but certainly not his admirer, employed 
spore characters (colour and septation) to divide the family into four genera, Ocellularia 
(colourless, transversely septate), Phaeotrema (brown, transversely septate), Thelotrema 
(colourless, muriform), and Leptotrema (brown, muriform). This system is easy to apply 
since spore characters are readily determined with a light microscope. It was reinforced when 
Zahlbruckner adopted it in his Catalogus Lichenum Universalis and Redinger (1936) 
followed it in his important study of the family in Brazil. Further subdivisions of the spore 
genera were founded on various macroscopic characters such as a degree of emergence, 
presence of a columella, etc., with little consistency, but as a result each genus contained 
essentially similar, parallel series of sections based on gross apothecial morphology. Species 
with identical apothecial structure were often as not artificially dispersed in different genera. 
This awkward classification deserves no further elaboration here; it is summarized by 
Redinger (1936) in considerable detail. 

Among recent lichenologists Salisbury (1971a) has vigorously challenged the continued 


THELOTREMATACEAE IN SRI LANKA 241 


use of spore genera in the Thelotremataceae. He rejects spores categorically and revives a 
classification based solely on excipular characters, recognizing one genus, Thelotrema, 
divided into three sections, Thelotrema (exciple colourless, with periphysoids), Myriotrema 
(exciple colourless or brownish, without periphysoids), and Ascidium (exciple carbonized, 
columella often present). These sections have the same circumscription as Fée’s three 
genera, Thelotrema, Myriotrema, and Ascidium (= Ocellularia). 

Although I used the spore genera in my previous studies of the family (Hale, 1974a, 
1978a), I have come to agree with Salisbury that excipular and other apothecial characters 
represent more natural traits than do spores. Indeed, spores, which so long served to separate 
genera, can be used to undermine them with equal cogency. For example, there are a number 
of species groups which have identical thallus morphology, excipular structure, and 
chemistry but spores of different colour or septation. A good case is the Thelotrema 
lepadinum-Phaeotrema lacteum-Leptotrema monosporum complex, three hypophloeodal 
species with a free, periphysoid-bearing exciple which manifestly associates them in one 
genus, 7helotrema. I had previously called these sporomorphs (Wirth & Hale, 1978: 2) in an 
attempt to rationalize their existence in spore genera. 

Another example is the reticulate-columellate series Ocellularia emersa-Phaeotrema 
virens-Leptotrema crassum, which have identical external morphology, carbonized exciple, 
and chemistry (psoromic acid). Their natural relationship is better expressed by placing 
them in one genus and regarding spore differences as a species character. Other examples will 
be cited below in the species list. They are not numerous, about 30 of the 450 species at the 
world level falling into this category, but more than adequately damage the credibility of 
spore-based genera. 

Having accepted excipular structures as a natural generic base, one soon realizes, however, 
that the use of one genus (Thelotrema) for 450 species is an unwieldy solution, especially 
since three well-delineated groups can be recognized. I have proposed to recognize Fée’s 
three genera for the reasons given below (see also Hale, 1980). 

The classic genus Thelotrema can be characterized by the constant presence of 
periphysoids and the frequent occurrence of a free (detached) exciple. Salisbury (1972a) 
assigned about 20 species here (as section Thelotrema) after studying some 200 type 
specimens. According to my own results, based on a study of over 350 types and much newly 
collected material, there are about 100 species of Thelotrema, of which 32 occur in Sri 
Lanka. 

A preliminary survey of the 100 species of Thelotrema brought out some interesting and, 
I believe, significant correlations with morphology and chemistry. For example, about 33 
species of Thelotrema at the world level lack a cortex, that is, are hypophloeodal. This 
suggests a high degree of primitiveness in the genus. The species are also distinctive in having 
a gaping pore, referred to as ‘apothecia excavata’ by Fée. In fact, species in Salisbury’s 
section Thelotrema, ‘platycarpum’ group, have an open disc surrounded by a conspicuous 
reflexed thalline rim, unique to the genus (Salisbury, 1972: 281). 

The chemistry of Thelotrema is particularly impoverished in comparison with 
Myriotrema and Ocellularia. Sixty of the species contain no lichen substances at all, a very 
high proportion compared to the other two genera. About 30 of the remaining species con- 
tain stictic acid and 14 others have fumarprotocetraric acid, hypoprotocetraric acid, proto- 
cetraric acid, norstictic acid, or salazinic acid. The most striking feature is the rarity of 
psoromic acid, otherwise the most common substance in the family. It occurs only in a 
small group of species, T. cryptotrema Nyl., T. piluliferum Tuck., T. platysporum Harm., 
T. saxicola Vainio, and T. sphinctrinellum Nyl., which do not, in any event, fit comfortably 
in the genus. 

Myriotrema, the second largest genus, is represented by nearly 150 species, 37 of them in 
Sri Lanka. Fée based it on immersed apothecia and the easily distinguished, thick cortex and 
medulla as seen in M. olivaceum. One of the unusual features of the genus is actually the high 
percentage of species with small immersed apothecia (0°2-0°4 mm in diameter) and small 
spores. Those with emergent or raised apothecia are classified by Salisbury in Thelotrema 


242 M. E. HALE 


section Myriotrema, the ‘T. bahianum’ group, which has an incurved thalline wall. The 
exciple in all the species is fused and colourless to brown, rarely lacking laterally. 

All species of Myriotrema have a distinct, often relatively thick medulla, and excepting M. 
glaucescens, M. multicava, and M. wightii, a distinct cortex, which 1s frequently aculeate and 
exfoliating as in Ocellularia. There are no hypophloeodal species. 

The chemistry of Myriotrema is nearly as varied as that of Ocellularia. Psoromic acid 
occurs in 50 (about 33%) of the 150 species known at the world level. Stictic acid is relatively 
more abundant (22 species) than in Ocellularia, but protocetraric acid (11 species) and 
hypoprotocetraric acid (4 species) less abundant. 

Ocellularia, treated by Salisbury as Thelotrema section Ascidium, is the largest genus in 
the family with about 200 species, of which 41 occur in Sri Lanka. It can be easily recognized 
by the fused, carbonized exciple, a trait which Montagne (1856: 364) used to characterize 
Ascidium, and the frequent occurrence of a central columella. While the apothecia vary 
greatly in size and emergence, the pore is usually constricted and distinct with an incurved 
thalline rim. Fée himself had emphasized the marginate, depressed pore when describing 
Ascidium. 

The cortical structure of Ocellularia is as complex as that in Myriotrema. Only about 10 
species lack a cortex. The corticate ones often have aculeate hyphae on the surface and 
exfoliation or splitting. 

The chemical profile of Ocellularia is very rich. There are many species, about 50 of the 
presently known 200, with psoromic acid, 25 with protocetraric acid, and 20 with 
hypoprotocetraric acid. Only about 50 species (25%) lack lichen substances, and only 11 
have stictic acid. When the species are divided into spore groups, the only noteworthy 
observation is that 12 species with transversely septate colourless spores have the 
‘cinchonarum’ unknown; only one muriform species has it. Otherwise the four spore types 
appear to have similar frequencies of each acid insofar as statistically significant numbers of 
species are compared. 


Family classification 


The family Thelotremataceae is readily identified by the pored apothecium. Santesson 
(1952: 306) briefly referred to the family in his exhaustive study of the foliicolous lichens. He 
felt that the family is closely related to the Graphidaceae, still a commonly held perception. 
Vézda (1966) studied a number of related genera and suggested affinities with the 
Diploschistaceae without making any formal proposals to combine them, this being done 
recently by Henssen & Jahns (1973: 374). Using the presence of periphysoids as a 
fundamental character, Vézda transferred Ramonia and Gloeolecta from the Gyalectaceae 
and placed them in the Thelotremataceae. Finally, Gilenstam (1969) places Thelotrema, 
Ocellularia, Phaeotrema, and Leptotrema along with the Graphidaceae, Diploschistes, and 
Conotrema under the order Ostropales without specifying a family breakdown. He pointed 
out a close relationship to Stictis, a periphysoid-bearing fungal genus, and a recent 
monograph on the ostropalean fungi by Sherwood (1977) confirms this. 

Zahlbruckner (1923) included Tremotylium Nyl. in the Thelotremataceae even though 
Vainio (1921: 184) had already discovered that 7. angolense Nyl., the type of the genus, is 
probably an Anthracothecium, since it has branched, anastomosing paraphyses. I would 
follow Santesson (1952) in transferring Gyrostomum, also placed in the Thelotremataceae 
by Zahlbruckner, to the Graphidaceae, and I would add Tremotylium sprucei Mill. Arg., 
which is not related to Tremotylium. The status of Polystroma Clemente, a very rare South 
American genus, is unclear at this time. 

The family Thelotremataceae as conceived now includes seven genera: Conotrema, 
Diploschistes, Gloeolecta, Myriotrema, Ocellularia, Ramonia and Thelotrema. It will almost 
certainly be expanded when more definitive studies on the Gyalectaceae are completed. 


y 


THELOTREMATACEAE IN SRI LANKA 243 
Chemistry 


As I earlier found in my studies of the Thelotremataceae in Dominica and Panama, the 
chemistry of this family is rich and diversified. The various species contain depsides and 
depsidones which can be determined easily with thin-layer chromatography [t.l.c.]. In my 
opinion, a knowledge of the chemical constituents is essential both for routine species 
identification and for a fuller understanding of speciation and evolution in the family. One 
would be seriously disadvantaged without this information. 

The lichen substances found in the 110 species from Sri Lanka are listed below 
alphabetically with the species in which they occur. All specimens were examined with t.l.c. 
using Merck F-254 precoated plates in two solvent systems (benzene-dioxane-acetic acid and 
hexane-ether-formic acid) and in a third (n-butanol-acetone-water) when distinguishing 
fumarprotocetraric acid and protocetraric acid. Plates were examined under shortwave ultra- 
violet, sprayed with 10% H,SO,, and heated at 100°C to develop the colours. Fuller des- 
criptions of problems in identification or the unknown compounds will be found in the list of 
species. 


Constictic acid (see stictic acid). 

Fumarprotocetraric acid: Myriotrema microstomum (with traces of protocetraric acid), M. minutum, 
Ocellularia thelotremoides (with nearly equal amounts of protocetraric acid), and Thelotrema 
nostalgicum. 

Hypoprotocetraric acid (usually with 4-0-demethylnotatic acid): Ocellularia eumorpha, O. 
kanneliyensis, O. neopertusariiformis, and O. triglyphica. 

Lichexanthone: Myriotrema fluorescens. 

Norpsoromic acid: See psoromic acid. 

Norstictic acid: Myriotrema compunctum (with stictic acid), M. porinaceum, and Thelotrema weberi. 

Protocetraric acid: Myriotrema microstomum (in traces with fumarprotocetraric acid), Ocellularia 
aurata, O. perforata, O. subsimilis, and O. thelotremoides (with equal amounts of fumarprotocetraric 
acid). 

Psoromic acid (often with norpsoromic acid): Myriotrema andamanicum, M. costaricense, M. 
frondosum, M. glaucophaenum, M. hartii, M. masonhalei, M. mastarion, M. microporum, 
M. minutulum, M. multicavum, M. rugiferum, M. terebratulum, Ocellularia crassa, O. emersa, O. 
exuta, O. fissa, O. lankaensis, O. marivelensis, O. pluripora, O. polillensis, O. rassagala, Thelotrema 
piluliferum, and T. platysporum. 

Salazinic acid: Ocellularia massalongoi. 

Stictic acid (often with constictic acid and associated unknowns A, and A,): Myriotrema 
anamalaiense, M. compunctum (with norstictic acid), M. decorticatum, M. desquamans, M. 
elachistoteron, M. eminens, M. fissurinum, M. glaucescens, M. nuwarense, M. santessonii, M. 
terebrans, Ocellularia melanotremata, O. pyrenuloides, O. sticticans, Thelotrema dilatatum, T. 
dissultum, T. novae-zelandiae, T. patens, T. phlyctidioides, T. platycarpoides, T. platycarpum, T. 
porinoides, and T. pulvereodiscum. 

No lichen substances present (anthraquinones may be present; see below): Myriotrema album, M. 
albocinctum, M. cinereoglaucescens, M. protoalbum, M. subconforme, M. wightii, Ocellularia 
albomaculata, O. ascidioidea, O. dolichotata, O. keralensis, O. leucomelaena, O. lirelliformis, O. 
meiosperma, O. meiospermoides, O. monosporoides, O. orthomastia, O. papillata, O. rhicnopora, O. 
tenuis, Thelotrema aggregatum, T. alborosellum, T. astroideum, T. colobicum, T. imperfectum, T. 
kamatii, T. lacteum, T. lepadinum, T. leprieurii, T. leprocarpoides, T. magnificum, T. monosporum, T. 
nureliyum, T. pidurutalagalum, T. pseudoexanthismocarpum, T. scabiomarginatum, T. subpatens, 


and T. waasii. ; 
K+ purple anthraquinones: Myriotrema wightii (medullary), Ocellularia punctulata (medullary), 


Thelotrema coccineum (epithelial), and 7. magnificum (epithelial). 

K+ yellow-orange pigments: Ocellularia aurata (epithelial). 

‘Cinchonarum’ unknown (P+ red-orange): Ocellularia neocavata, O. punctulata. 

‘Chonestoma’ unknown (with upper ‘olivacea’ unknown) (P—): Ocellularia chonestoma, O. 
croceopora, O. nylanderiana. 

‘Diacida’ unknown (P+ yellow-orange): Ocellularia diacida. 

‘Leightonii’ unknown (P+ orange-red): Myriotrema granulosum. 

‘Olivacea’ unknown (P—): Myriotrema olivaceum. 


244 M. E. HALE 


‘Pertusariiformis’ unknown (P—): Ocellularia pertusariiformis. 
‘Platysporum’ unknown (P+ red): Thelotrema platysporum. 
‘Thwaitesii’ unknown (P+ orange): Myriotrema polytretum, M. thwaitesii. 


The chemistry of the family in Sri Lanka is rather close to that in the New World. The four 
commonest profiles have approximately the same frequency: no substances present (34% of 
the species in Sri Lanka vs 23% in Panama and Dominica), psoromic acid (21% vs 20%), 
stictic acid (18% vs 16%), and protocetraric acid (4% vs 8%). None of the P+ unknowns from 
Sri Lanka (except for the ‘cinchonarum’ unknown) nor salazinic acid have yet been 
discovered in the New World. On the other hand, one P+ unknown, the ‘praestans’ 
unknown, occurs in the New World but not in the Old (Hale 1974a: 36). 


Ecology and conservation 


As one would expect, the majority of the species in Sri Lanka occur in relatively narrow 
ecological niches, and for this reason information on habitat preferences can be invaluable in 
distinguishing the species. The commoner ones are discussed below. The less commonly 
collected species (e.g. those collected less than 6-8 times) unfortunately cannot be 
categorized with any degree of confidence and are omitted in the discussions; their habitats 
are briefly described in the main list of species. 

Elevation has the greatest effect on the overall distribution of the Thelotremataceae. 
Naturally, elevation extremes are limited by the particular topography of a country and 
assignment of species to high or low elevation classes should be taken in a relative sense. For 
example, Thelotrema weberi occurs at the highest elevation in Sri Lanka (2520 m) but was 
first collected in New Guinea at 4450 m. For Sri Lanka, I define low elevation as terrain 
lying between sea level and about 350 m. High elevation is 2000 m and above. The few 
localities at about 1000 m (Enselwatta and Rassagala) have both high elevation and low 
elevation species although low elevation species predominate. 

The high elevation species, mostly in the genus Thelotrema, were collected chiefly in the 
Nuwara Eliya district (Horton Plains), Pidurutalagala, and the Hakgala Botanical Garden. 
They include Myriotrema nuwarense, Ocellularia monosporoides, Thelotrema lepadinum, 
T. nostalgicum, T. nureliyum, T. pidurutalagalum, and T. weberi. A division of the species 
into base level and canopy types at this elevation is meaningless since the trees are rather low 
and form a broken canopy without strong vertical zonation of the epiphytes. 

Low elevation species grow chiefly in virgin forests below about 350 m but sometimes 
occur as high as 850-1000 m. In Sri Lanka, these include Myriotrema albocinctum, M. 
album, M. desquamans, M. granulosum, M. olivaceum, M. terebratulum, Ocellularia 
aurata, O. chonestoma, O. crassa, O. croceopora, O. dolichotata, O. keralensis, O. 
lankaensis, O. marivelensis, O. massalongoi, O. melanotremata, O. nylanderiana, O. 
papillata, O. perforata, O. polillensis, O. punctulata, O. thelotremoides, Thelotrema 
magnificum, T. platysporum, and T. scabiomarginatum. 

It is possible to classify many of the species further as being restricted to either the canopy 
or base level of the trunk because I was able to collect in the canopy of felled trees at logging 
sites. Vertical zonation of lichens is conspicuous in mature rain forests. The commonest 
canopy species are Myriotrema album, M. eminens, M. microporum, M. olivaceum, M. 
terebratulum, Ocellularia aurata, O. exuta, Thelotrema piluliferum, and T. platysporum. 
The preponderance of Myriotrema species is remarkable. 

The largest group of species was restricted to the base level on exposed roots, buttresses, 
saplings, and the trunks up to 1-3 m in dense forest: Myriotrema fluorescens, M. micro- 
stomum, M. thwaitesii, Ocellularia chonestoma, O. croceopora, O. dolichotata, O. 
lankaensis, O. massalongoi, O. melanotremata, O. papillata, O. perforata, O. thelotre- 
moides, Thelotrema alborosellum, T. astroideum, T. colobicum, T. magnificum, T. 
porinoides, and T. scabiomarginatum. The remaining species in the three genera are not 
strongly restricted to any particular level in the forest. 


THELOTREMATACEAE IN SRI LANKA 245 


One of my goals in Sri Lanka was to assess the survival of the lichen flora in the face of 
habitat destruction first noticed by Thwaites (1864). It is most reassuring that I was able to 
recollect virtually all of the species which Thwaites found both in lowland rain forest and at 
higher elevations. One can conclude that the continuity of forest cover is still adequate to 
maintain a lichen population that we know to be fatally dependent on virgin rain forest. 

In the last 120 years since Thwaites collected in Sri Lanka, however, there have been great 
and irreversible changes in the vegetation. Virtually all of the accessible forests along the 
west coast have been cut down for rubber plantations, general agriculture, and expansion of 
villages. At higher elevations (over 800 m) huge areas of rain forest were cut for tea estates 
even in Thwaites’ time. 

Only one large tract of more or less unbroken forest remains, the Sinharaja Forest Reserve, 
an area about 220 km? of great importance as a watershed. Even this forest is being destroyed 
by commercial logging, leaving behind a desolate landscape subject to erosion and at best 
revegetated by secondary scrub and bamboos. Logging is particularly damaging for lichens 
and many other organisms, since exposure to direct sunshine and desiccation can wipe them 
out in a matter of months after the canopy is destroyed. One can see many bleached thalli on 
trees still standing in logged areas. Species on felled trees that are not removed in a logging 
operation die off because of this changed environment, burning or overgrowth by weedy 
scrub. 


Phytogeography 


Few world lichen monographs have been based on adequate collections, and any discussions 
of distribution are severely limited by the poor geographic coverage. Every recent expedition 
by lichenologists to tropical countries has turned up totally unexpected range extensions. 
This is truer perhaps for the Thelotremataceae than for almost any other crustose family 
because they often cannot be recognized in the field without time-consuming search with a 
hand lens. I will try, however, to summarize very briefly the thelotreme flora and the 
geographic patterns that are slowly emerging for the family, basing my statements on 
specimens which I have personally verified. 


The flora 


The thelotreme flora of Sri Lanka based on collections made by Thwaites, Almquist, Alston, 
Wheeler and Fosberg, Moberg and Santesson comprised 36 species amongst 200 specimens 
(of which about 110 are Thwaites duplicates). My own excursions in 1976 and 1978 added 
about 715 specimens and 74 species, bringing the total of 110 species and about 915 speci- 
mens on which the present revision has been based. 

Ocellularia is by far the largest genus with 41 species and 414 collections. The 10 
commonest species are O. punctulata (46 collections), O. chonestoma (43), O. papillata (37), 
O. lankaensis (36), O. thelotremoides (24), O. crassa (18), O. keralensis (18), O. 
nylanderiana (17), O. polillensis (15), and O. perforata (14). Five species, O. diacida, O. 
leucomelaena, O. neocavata, O. rassagala, and O. tenuis, were collected only once, while the 
remaining 26 species were each collected between 2 and 12 times. 

Myriotrema is the second most common genus, represented by 37 species and 224 
collections. The commonest species, M. olivaceum, was collected 28 times. Myriotrema 
desquamans is represented by 23 collections, M. rugiferum by 23, M. masonhalei by 22, M. 
microstomum by 13, and M. albocinctum by 11. The remaining 31 species are much rarer 
with | to 9 collections apiece. 

Thelotrema ranks close to Myriotrema with 32 species and 188 collections in Sri Lanka. 
The five commonest species account for 107 of these collections: 7. imperfectum (37), T. 
platysporum (23), T. scabiomarginatum (18), T. porinoides (15), and T. magnificum (14). 
Most of the species (22) were collected 3 or fewer times, the remainder 4-11 times. 

When we analyze the percentage of the flora according to genus, we find that Thelotrema 
has 29% of the species, Myriotrema 34%, and Ocellularia 37%. Panama has relatively fewer 


246 M. E. HALE 


species of Thelotrema (21% of the flora), more Ocellulariae (44%), and the same number of 
Mpyriotremata (35%). 


Phytogeographic affinities 


The pantropical species in Sri Lanka may be defined very broadly as those which occur in 
the New World and in one or more tropical Asian countries or also in Africa, a poorly 
collected region, and in Australia. Using this definition, we can identify the following 34 
pantropical species in Sri Lanka, about one-third of the total 110 in the flora: Myriotrema 
album, M. compunctum, M. costaricense, M. glaucescens, M. glaucophaenum, M. hartii, M. 
minutulum, M. olivaceum, M. terebratulum, M. wightii, Ocellularia aurata, O. emersa, O. 
fissa, O. leucomelaena, O. lirelliformis, O. meiosperma, O. papillata, O. perforata, O. 
pyrenuloides, O. subsimilis, O. tenuis, O. triglyphica, Thelotrema aggregatum, T. 
alborosellum, T. coccineum, T. dilatatum, T. lacteum, T. lepadinum, T. leprieurii, T. 
monosporum, T. phlyctidioides, T. platycarpoides, T. platycarpum, and T. porinoides. The 
species are more or less equally divided among the three genera. 

A second group of 37 species occurs in Sri Lanka and in one or more Asian countries but 
not in the New World: Myriotrema anamalaiense, M. andamanicum, M. cinereo- 
glaucescens, M. desquamans, M. eminens, M. fissurinum, M. masonhalei, M. microporum, 
M. microstomum, M. minutum, M. polytretum, M. porinaceum, M. rugiferum, M. 
santessonil, M. subconforme, M. terebrans, Ocellularia chonestoma, O. crassa, O. diacida, 
O. dolichotata, O. eumorpha, O. keralensis, O. marivelensis, O. massalongoi, O. mono- 
sporoides, O. nylanderiana, O. orthomastia, O. polillensis, O. thelotremoides, Thelotrema 
colobicum, T. kamatii, T. novae-zelandiae, T. patens, T. pseudoexanthismocarpum, T. 
piluliferum, T. platysporum, and T. weberi. The relatively large number of Myriotrema 
species and the small number of Thelotrema species is noteworthy when compared with the 
pantropical listing. 

We should add to the Asian element the following 39 species which are known only from 
Sri Lanka. Some of these will eventually be found in the dipterocarp forests of south-east 
Asia: Myriotrema albocinctum, M. decorticatum, M. elachistoteron, M. fluorescens, M. 
frondosum, M. granulosum, M. mastarion, M. multicavum, M. nuwarense, M. protoalbum, 
M. thwaitesii, Ocellularia albomaculata, O. ascidioidea, O. croceopora, O. exuta, O. 
kanneliyensis, O. lankaensis, O. meiospermoides, O. melanotremata, O. neocavata, 
O. neopertusariiformis, O. pertusariiformis, O. pluripora, O. punctulata, O. rassagala, O. 
rhicnopora, O. Ssticticans, Thelotrema astroideum, T. dissultum, T. imperfectum, T. 
leprocarpoides, T. magnificum, T. nostalgicum, T. nureliyum, T. pidurutalagalum, T. 
pulvereodiscum, T. scabiomarginatum, T. subpatens, and T. waasii. 

The 110 species in Sri Lanka fall into three almost equal groups, the pantropical element 
(34 species), the Asian element (37 species), and the endemics (39). By comparison, Panama 
(Hale, 1978a), represented by a flora of comparable size (99 species) and collecting intensity, 
has 28 pantropical species, 46 New World species, and 25 endemics. Of the pantropical 
element, Panama and Sri Lanka share only 19 species. 

A further breakdown of the species in the Old World element is difficult without access to 
more collections. However, a number of species in Sri Lanka seem to be range extensions 
from the dipterocarp forests of the Philippine-Malaysian region (e.g. Myriotrema 
cinereoglaucescens, M. minutum, M. polytretum, Ocellularia chonestoma, O. crassa, O. 
dolichotata, O. marivelensis, O. nylanderiana, O. orthomastia, O. polillensis, O. triglyphica, 
and Thelotrema patens) and these make up the most conspicuous phytogeographic element 
in the flora. Considering its proximity, India has surprisingly little in common with Sri 
Lanka. The intensively collected Western Ghats with about 50 species and Sri Lanka with 
110 share only 8 endemics: Myriotrema anamalaiense, M. fissurinum, M. masonhalei, 
Ocellularia diacida, O. keralensis, O. thelotremoides, Thelotrema kamatii and T. 
pseudoexanthismocarpum. Few if any of the 39 endemic species of Sri Lanka can be 
expected in India. 


THELOTREMATACEAE IN SRI LANKA 247 


As I had discovered in Panama and the West Indies, the Thelotremataceae are most 
abundant in undisturbed, low elevation rain forest, and in Asia this is the ideal habitat for the 
Dipterocarpaceae. Trees in this family make up much of the forest cover below 1000 m 
elevation from Sri Lanka to the Philippines and Solomon Islands. We cannot say yet which 
of the major tropical forest biomes have the greater number of species, those in the New 
World or in Asia; perhaps they are equally rich. It cannot be denied, however, that the family 
has reached its present state of evolution and species diversity in tropical rain forests. I would 
not agree with Salisbury (1975: 59) that the Thelotremataceae is a temperate family which 
has migrated to the tropics. This may be true of Thelotrema but not of Myriotrema or 
Ocellularia. 


Collecting localities 


While Sri Lanka is a rather small land mass, about three-fourths the area of Austria, it 
embraces a very great range of land forms and vegetation types. About 10% of the island 
in the lower south-west portion is dense mountainous virgin rain forest. Moving to the east 
and north the land is flatter, quite dry, and very much disturbed. It is obvious that Sri Lanka 
is far from thoroughly collected and a complete study would take many years. Most 
collectors in the past have gone to the Nuwara Eliya-Horton Plains area, a high cool plateau 
not really representative of the forested area as a whole. Southwest of this plateau stretches 
the Sinharaja Forest Reserve, Kanneliya Forest Reserve, Gilimale Forest Reserve, and 
adjacent rain forest at 100-350 m elevation, little explored for lichens and now being logged 
selectively for Dipterocarpaceae. It is a region of high humidity, continuous heat, and 
abundant land leeches. This is where I concentrated my collecting efforts, taking advantage 
of access to the forest along freshly cut logging roads. 

Unfortunately, time did not permit excursions to the ‘Knuckles’ east of Adam’s Peak, the 
Badulla District, which must have an interesting, though probably less rich thelotreme flora. 
Nor did I visit the semi-arid scrub where several interesting species occur. 

To conserve space in the species list, I have abbreviated my collecting localities as follows. 
All originals are in US with numerous duplicates in BM and PDA. The Thwaites collections 
and those by later workers are cited in full with the herbarium indicated by standard 
acronyms. My collection numbers in the 40 000 series were made in the trip of March 1976, 
those in the SO 000 series in February 1978. 

1. Maliboda to Theberton Estate, Kegalla District. This is a trail near an electric 
transmission line at about 800 m elevation, crossing the western end of the Adam’s Peak 
Reserve and overlooking Norton Bridge to the north. The forest is still largely virgin in the 
upper region. 

2. Rassagala, Ratnapura District. The trail climbs to a long ridge off the Halwathura- 
Kanda road northwest of Balangoda at 850-900 m elevation. It is entirely virgin rain forest 
with mossy forest on the highest ridge. While there was no logging here, the rather low trees 
could be reached easily. 

3a. Gilimale, Ratnapura District. This is a hilly forested region at 150-200 m elevation, 
the Gilimale Forest Reserve, now being logged with elephants. The canopy flora of the 
dipterocarp trees was sampled extensively. 

3b. IBP Reserve, Ratnapura District. I sampled this virgin dipterocarp forest only briefly. 
It lies on a steep slope below Adam’s Peak at about 200 m elevation. 

4. Weddagala, Ratnapura District. This is a very large area of virgin rain forest in the 
Sinharaja Forest Reserve at about 300 m elevation. Mechanized logging was in operation 
during my first visit in 1976 but had ended by 1978. I collected from a network of logging 
roads over an area of perhaps 6 km? at 6 sites. Large samples were obtained from the canopy 
and from lower levels in undisturbed forest. This is the best collected area. 

5. Enselwatta, Ratnapura District. This is a plateau above Deniyaya at about 1100 m 
elevation, a mossy forest well known to phanerogamic collectors. It is similar to Rassagé 
but a little more exposed. 


248 M. E. HALE 


6. Hedigalla, Kalutara District. This is a virgin rain forest at about 200 m elevation east of 
Matugama. It is being logged actively. 

7a and 7b. Morapitiya, Kalutara District. This is a continuation of the Hedigalla forest 
south of Badureliya. One site (7a) was called Ruhunu-Kanda, a dense rain forest at about 
200 m elevation being logged now. The other site (7b) is Ambalam-pola at closer to 300 m 
elevation, now being logged with elephants. It includes some collections from a new 
roadhead about | km farther south. Freshly felled dipterocarp trees were available at these 
sites. 

8. Morawaka, Matara District. This is a low mountain pass of not more than 400 m 
elevation near Morawaka which we traversed on an old trail. It is barely disturbed 
dipterocarp forest not yet logged. 

9. Hiniduma, Galle District. There is an extensive area of lowland rain forest at 
100-150 m elevation southeast of Hiniduma, called the Kanneliya Forest Reserve. It was 
being logged with machinery in 1976 and had previously been heavily logged. 

10. Nuwara Eliya to Hakgala, Nuwara Eliya District. This site is a forest reserve about 
1 km off the road between Nuwara Eliya and Hakgala Botanical Garden. It is a rather 
sheltered forest at about 2100 m elevation. 

11. Hakgala Botanical Garden, Nuwara Eliya District. Thwaites collected lichens in this 
well known garden. The large trees are all planted. The elevation is near 2100 m. 

12. Pidurutalagala, Nuwara Eliya District. This long, gently sloped mountain is the 
highest point in Sri Lanka (8260 ft. or 2518 m), reached by a well-worn trail. The forest 
along the trail has been severely disturbed since Thwaites’ time. The summit is covered with 
a dense almost impenetrable shrub draped with Usnea. Most of the lichens are found on 
exposed branches. 

13. Peradeniya Botanical Garden, Kandy District. I made a few collections on trees in this 
long-established garden. The elevation is about 400 m. 

The richest localities were the virgin lowland rain forests. I found 44 different species in 
the Weddagala-Sinharaja area. There were 43 species in the Hiniduma-Kanneliya Forest 
Reserve and 42 in the Morapittya area, all lying at 150-300 m elevation. By contrast only 15 
species have been collected in the Nuwara Eliya region (elevation about 2000 m), one of the 
most frequently visited localities in Sri Lanka. There were only eight species at 
Pidurutalagala, the highest peak. 


List of species 


The 110 species of the family Thelotremataceae in Sri Lanka are listed below in alphabetic 
order under three genera, Thelotrema, Myriotrema, and Ocellularia. The synonymy includes 
species from both the New World and Old World, but I have not repeated data on synonymy 
previously published in my studies of the family in Dominica and Panama (Hale, 1974a, 
1978a). These studies should also be consulted for additional comments on the species. 

In compiling the descriptions, I have taken anatomical measurements of the cortex, algal 
layer, and medulla from scanning electron microscope prints and free-hand light microscope 
sections. Apothecial and pore size was measured from photographs reproduced at x 14 
magnification. The exciple and periphysoids are best seen in sections mounted in Hoyers 
Solution or other clearing agents; water mounts are not satisfactory. 

Chemistry is an important and useful character when studying the Thelotremataceae. 
While colour tests are often adequate, identification of the substances with thin-layer 
chromatography is desirable. Chemical constituents given in the species descriptions are 
those of the type material and the specimens listed. 

I did not have suitable Sri Lankan material available for illustrating Thelotrema lacteum 
and 7. monosporum. Collections from India are substituted for these. Localities for Hale 
collections are abbreviated 1-13 as indicated above under ‘Collecting localities’ (p. 247). 


_ 


THELOTREMATACEAE IN SRI LANKA 249 


Key to the genera 


1 _Exciple with periphysoids on the inner face; columella always lacking Thelotrema (p. 249) 

—  Periphysoids lacking; columella present orabsent. .  . 2 

2  Exciple colourless to pale reddish brown; columnar structures, if Present ‘Acnéacbodined 
Myriotrema (p. 271) 

- Beenie Pi ionized: pi thanived columella Gian picsent. i : ‘ Ocellularia (p. 297) 


I. THELOTREMA Ach. 
Meth. Lich.: 130 (1803). 


Antrocarpum G. Meyer, Nebenstund. : 326 (1825). 

Gomphospora Trevisan in Riy. Period. Accad. Padova 1851-52 : 268 (1852). 
Brassia Massal. in Atti R. Ist. veneto Sci. U1, 5 : 259 (1860). 

Asteristion Leighton in Trans. Linn. Soc. Lond. 29 : 163 (1870). 
Chroodiscus (Mill. Arg.) Miill. Arg., Lich. Epiphyll. Novi: 18 (1890). 


Thallus crustose, epiphloeodal or hypophloeodal; cortex when present loosely organized to 
dense, rarely splitting and exfoliating; medulla usually rather thin or lacking; apothecia 
somewhat immersed to generally raised or strongly emergent, the thalline wall incurved to 
erect to recurved; exciple free (detached) or rarely fused (attached), the inner face covered 
with short periphysoids; pore often broad and gaping; spores colourless or brown, 
transversely septate or muriform, I+ blue or I—. 


TYPE SPECIES. Thelotrema lepadinum (Ach.) Ach. 
NUMBER OF SPECIES. About 100. 


OBSERVATIONS. 7helotrema is amply characterized by the presence of periphysoids. There 
is great variability in emergence, from nearly immersed and flush (e.g. 7. kamatii) to 
emergent (7. nostalgicum). Salisbury (1972a) divides Thelotrema sect. Thelotrema into two 
groups, the ‘7. /epadinum’ group (thalline wall incurved) and the ‘7. platycarpum’ group 
(thalline wall recurved). While extreme examples of these groups such as T. /epadinum (Fig. 
8a) and T. platycarpum (Fig. 9f) are easily separated, there are a number of species with an 
erect thalline wall and a relatively open disc (e.g. 7. aggregatum, T. scabiomarginatum) 
which are difficult to place. There is also a third small but distinctive group with emergent 
apothecia, a fused exciple, and psoromic acid, including T. piluliferum and T. platysporum 
from Sri Lanka, which do not fit in any of these categories. I feel it is premature to divide the 
genus into groups or formal taxa as long as so few ontogenetic studies have been made and a 
significant number of ‘intermediates’ seems to exist. 


Key to the species 

1 Spores colourless . 2 
-— Spores brown ; ; ; ; ; ; : : : : 26 
2 Spores transversely septate ‘ , : : ; : 5 , , 2 , f ; 3 
— Spores muriform . : ; ; : : ; ' : : ; 17 
3 __ Disc open, brightly pigmented purplish to orange : 4 
— Discopenorclosed,not pigmented. . ee eee ee ee 6 
4 Collected on evergreen leaves of trees in rain forest ; : F F 4. T. coccineum 

— Collected on tree bark . : 5 : 4 anew = 5 
5 Cortex lacking; disc deep purple: spores 152 18 um 2 Oia ene0 31. T. waasii 

— Cortex present; disc orange; spores 70-90 um MP yw >. Se 14. T. magnificum 

6 Spores large, 75-250 um long, 12-30 loculate ; , : ; ; : : f 
— Spores small to medium sized, 10-65 wm long, 4-12 loculate ; 12 
7 


Apothecia strongly emergent and urceolate; thallus K—, P+ red (fumarprotocetraric acid) 
Lhe ne k sal, st 16. T. nostalgicum 


250 


M. E. HALE 
Apothecia immersed to flush or semi-emergent; thallus K+ yellow, P+ orange (stictic acid) or 
K-—, P—(no substances present) : . ee sees ene : , 26. 
Cone ilkewineecnciea@iclrenam 5 56 o .6 6 oo « 26. i Ne porinoides 


Cortex present; no substances present ; 
Apothecia 1-1-5 mm diam with erupting excipular material 
27. T. pseudoexanthismocarpum 
Apothecia 0°4-0°8 mm diam; excipular material not erupting 
Apothecia more or less emergent; thalline rim incurved to suberect: spores less than 70 um 


long , SF tes 8. T. imperfectum 
Apothecia immersed to flush: thalline rim barely raised; spores more than 100um long 

Spores 120-150 wm long; collected atlowerelevations. . . . . 9. T. kamatii 
Spores 60-120 um long; collected at higher elevations . , ; ; 18. T. nureliyum 


Disc open, 1 mm wide or more; thalline rim distinctly recurved . 
Disc more or less closed; thalline rim low, suberect to barely recurved 


Dhallusiconticates darkolive greent) - 1.) n-ne 3. T. astroideum 
Thallus lacking a cortex, whitish gray SOC) aR. gee ee ees ee et re 
Thallus K+ yellow, P+ orange (sticticacid) . ; ; : : : 6. T. dilatatum 
Thallus K-, P— (no substances present) ‘ ' 2. T. alborosellum 
Apothecia semi-emergent; thalline rim incurved; spores 35-70 um : 8. T. imperfectum 
Apothecia more or less flush; thalline rim low, becoming erect; spores 12-26 um . 

Thalline rim coarse; thallus K+ yellow, P+ orange (stictic acid) . 20. T. phlyctidioides 
Thalline rim delicate; thallus K-,P—(no substances) . , : . 30. T. subpatens 


Spores large, over 100 zm long 
Spores smaller, less than 80 wm long “is 5" gecr 1 ftoee SL eh Sona oe 
Apothecia strongly emergent,urceolate . =. sw sk 32. T. weberi 
Apothecia flush to semi-emergent . 
Thallus shiny, distinctly corticate; disc filled with excipular material 

29. T. scabiomarginatum 
Thallus'dull; lacking‘a. cortex;diseopen’ = :° CC Se ee 
Thallus K+ yellow, P+orange(sticticacid) . . . . . . . 19. T. patens 
Thallus K-, P—(no substances present) . 5. T. colobicum 
Pore very tiny, 0:05-0:15 mm diam; thallus P+ deep yellow (psoromic acid) 
Pore gaping, 0-2-0-4 mm diam or disc open, chroodiscoid : 7 , : 
Apothecia strongly emergent, 0°7—-1:0 mm diam; pore 0:05 mm diam 22. T. piluliferum 
Apothecia raised to emergent, 0:3-0:07 mm diam; pore variable, to 0:15 mm diam 2 Sale 


platysporum 
Thallus corticate, dark olive green. dey Selon 7. T. dissultum 
Thallus lacking a cortex, dull, whitish to straw ‘coloured eT «ct hes-yyt chp ene 
Apothecia emergent; thalline rim incurved with a distinct pore . : 11. T. lepadinum 
Apothecia flush to barely raised; thalline rim erect to recurved, the disc open . : 
Thallus K+ yellow, P+ orange(sticticacid) . . . . . Ave novae-zelandiae 
Thallus K-, P—(no substances present) . : 3s We me 
Spores transversely septate (rarely witha longitudinal septum in end ae 
Spores muriform . fe We an, 9 ES, 
Spores large, 90—100um; apothecia urceolate - ; 2 l. ae pidurutalagalum 
Spores 15-60 wm long; apothecia immersed to emergent, not urceolate 
Apothecia distinctly pored with anincurvedthalline wall . . . 10. T. lacteum 
Apothecian with an open disc, chroodiscoid; thalline wall erect to recurved errr 
Thalline rim thick, pulverulenttoflocculent. . . . . . 12. T. leprieurii 


Thalline rim thin, often recurved, entire to white pruinose . ae es 
Disc 0:S-O0'7 mm diam, becoming filled with excipular material: thallus K-, P- (no 


substances present) 1. T. aggregatum 
Disc 0:5-2°5 mm diam, open; thallus K+ yellow, P+ orange (stictic acid) eg 2 

Disc 0°5—1°2 mm diam ‘ ; ; é : : : : : 23. T. platycarpoides 
Disc 1-2°5 mmdiam . F : 24. T. platycarpum 
Apothecia chroodiscoid; thalline rim ‘erect, pulverulent ; : 28. T. pulvereodiscum 


Apothecia with a distinct pore, the disc closed; thalline rim incurved 15. T. monosporum 


10 


11 


THELOTREMATACEAE IN SRI LANKA 251 


1. Thelotrema aggregatum (Hale) Hale in Mycotaxon 11 : 131 (1980). 
(Fig. 7a) 


Phaeotrema aggregatum Hale in Smithson. Contr. Bot. 16:29 (1974) Type: Dominica, 
Newfoundland, January 1969, M. E. Hale 35 229 (US—holotype). 
Icones. Hale, 1974a : 26 fig. 13h (holotype). 


Thallus olive greenish, 6-10 cm broad, shiny, waxy, continuous; cortex dense, 15-18 wm, 
with sparse aculeate hyphae, some internal splitting; algal layer continuous, 154m; medulla 
10-40 um with crystals and periderm inclusions; apothecia more or less immersed, solitary 
to aggregated in dense clusters, 0°5-O'7 mm diam, the thalline rim barely discernible to 
emergent, suberect or even rarely weakly recurved, the tips granular and pruinose, the 
exciple free, apically reddish, partially filling the disc; disc dark or white pruinose; 
hymenium 70-90 um; spores brown, transversely septate, 5-7 x 12-18 um, 4-6 loculate, I+. 


CHEMISTRY. No substances present. 
HasitTaT. Canopy branches in rain forest at lower elevations (150-300 m). 
DISTRIBUTION. Dominica, Sri Lanka. 


ADDITIONAL SPECIMENS. Hale collections: 3a (46 345), 7b (51 021). Thwaites collections: 
C.L. 167 (BM, PDA, S, UPS) (as ‘Lgt. 84. Thelotrema auberianum’). 


OBSERVATIONS. This species is probably assignable to the ‘7helotrema platycarpum’ group, 
most closely related to T. leprieurii (see below), from which it can be distinguished by the 
smaller, aggregated apothecia. Leighton called the Thwaites material ‘7. auberianum Mont.’ 
but this reticulate-columellate species (= Ocellularia auberiana (Mont.) Hale; see Hale, 
1978a : 35) contains psoromic acid. Nylander had not verified this collection. 


2. Thelotrema alborosellum (Nyl.) Tuck., Gen. Lich. : 139 (1872). 
(Fig. 7b) 


Graphis alborosella Nyl. in Annls Sci. nat. (Bot.) IV, 19 : 372 (1863). Type: Colombia, Lindig 2694 
(H-Nyl.—lectotype; BM, FH-Tuck., UPS—isolectotypes). 

Ocellularia alborosella (Ny|.) Santesson in Symb. bot. upsal. 12(1) : 308 (1952). 

For additional synonymy see Hale (1978a: 12). 

Icongs. Hale, 1974a: 18 fig. 9b (lectotype).—Hale, 1978a : 13 fig. 4c. 


Thallus whitish, dull, to 10 cm diam; cortex lacking, the algae scattered among superficial 
medullary hyphae; medulla mostly hypophloeodal; apothecia chroodiscoid, to 1 mm diam, 
the thalline rim split and recurved, the exciple fused to barely free apically, colourless; disc 
tannish to whitish pruinose; hymenium 60-75 wm; spores colourless, transversely septate, 
4-5 x 12-24 um, 6-8 loculate, I—. 


CHEMISTRY. No substances present. 
HasitTaT. Lower bole in rain forest or disturbed sites at 200-2100 m elevation. 
DISTRIBUTION. United States, West Indies, Central and South America, India, Sri Lanka. 


ADDITIONAL SPECIMENS. Hale collections: 1 (50209), 2 (51 122, 51 236), 4 (50 426), 10 (50 
269). 


OBSERVATIONS. This species is quite variable (see discussions under Ocellularia alborosella 
in Hale, 1974a:16 and 1978a:12) and my identifications encompass a rather broad 
population. The type specimen (and material from Sri Lanka) definitely lacks a cortex. 
Nylander gives spore size as S—6 x 20-23 wm and my collections from Sr Lanka fall in this 
range. 


252 M. E. HALE 


Fig. 7 Species of Thelotrema. (a) T. aggregatum (Hale, 51 021). (b) 7. alborosellum (Hale 50 269). 
(c) T. astroideum (Hale 47 035). (d) T. coccineum (Thwaites s.n. in BM). (e) T. colobicum (Hale 
50 251). (f) T. dilatatum (Hale 46 218). (g) T. dissultum (Hale 51 139). (h) T. imperfectum (Hale 
50 433). (1) T. kamatii (Hale 50 098). Scale in Fig. 71= 1 mm. 


THELOTREMATACEAE IN SRI LANKA 253 


3. Thelotrema astroideum (Berk. & Broome) Hale in Mycotaxon 11 : 131 (1980). 
(Fig. 7c) 


Platygrapha astroidea Berk. & Broome in J. Linn. Soc. (Bot.) 14: 109 (1875). Type: Sri Lanka, 
Thwaites 629 (K—lectotype). 

Ocellularia astroidea (Berk. & Broome) Hale in Mycotaxon7 : 377 (1978). 

Icones. Hale 19785 : 384 fig. 1. 


Thallus dark greenish or brownish gray, 6-15 cm broad, shiny, continuous; cortex dense, 
10-20 wm; algal layer 10-15 wm; medulla not distinct, mostly hypophloeodal; apothecia 
chroodiscoid, 0°8—1:2 mm diam, the thalline rim split and recurved, the exciple fused with 
short periphysoids; disc and inner surface of thalline wall white pruinose; hymenium 
40-65 um; spores colourless, transversely septate, 3-5 x 10-18 wm, 4-6 loculate, I—. 


CHEMISTRY. No substances present. 
HasitTat. Base and lower bole, saplings in rain forest at low to mid elevations (150-900 m). 
DISTRIBUTION. Sri Lanka. 


ADDITIONAL SPECIMENS. Hale collections: 2 (51 157,51 177,51 199, 51 203), 3 a (46 297, 46 
365), 4 (47 127), 7a (50 164), 7b (51 103), 8(47 037). 


OBSERVATIONS. The contrast between the dark waxy greenish thallus and the large white 
apothecia make this a conspicuous species. It can be differentiated from T. al/borosellum by 
the presence of a cortex as well as by the more delicate, strongly recurved thalline margin. It 
is known only from Sri Lanka (Hale, 19785 : 377). 


4. Thelotrema coccineum (Leighton) Hale in Mycotaxon 11 : 131 (1980). 
(Fig. 7d) 


Platygrapha coccinea Leighton in Trans. Linn. Soc. Lond. 25 : 456 (1866). Type: Brazil, Spruce (not 
seen). 

Ocellularia coccinea (Leighton) Mill. Arg. in Flora, Jena 66 : 353 (1883). 

Chroodiscus coccineus (Leighton) Miill. Arg., Lich. Epiphyll. Novi: 18 (1890). 

See Santesson (1952 : 309) for full synonymy. 


Thallus foliicolous, greenish, 2 cm broad; cortex lacking; apothecia chroodiscoid, barely 
erumpent, 0°2-0'4 mm diam, the thalline rim recurved, exciple free; disc plane, scarlet; 
hymenium 35-60 um; spores colourless, transversely septate, 2-4 x 8-12 wm, 2-4 loculate 
(spore data from Santesson). 


CHEMISTRY. K+ purple epithecial pigment. 
Hasitar,. Palm leaves. 

DISTRIBUTION. Pantropical. 

SPECIMEN. Sri Lanka, Thwaites s.n. (BM). 


OBSERVATIONS. Santesson recognized this epiphyllous species as Chroodiscus coccineus 
(Leighton) Miill. Arg. with some hesitation since the thin-walled spores are not basically 
different from those in other chroodiscoid species. I am listing it under Thelotrema. 


5. Thelotrema colobicum Nyl. in Bull. Soc. Linn. Normandie Il,7 : 169 (1873). 
(Fig. 7e) 


Type: India, Andaman Islands, 1867, Kurz (H-Nyl. 22493—lectotype). 
Thelotrema deightonii C. Dodge, Beth. nov. Hedwigia 12 : 100 (1964). Type: Sierra Leone, Kori, Njala, 
Deighton M5856 (BM—holotype). 


254 M. E. HALE 


Thallus whitish gray, 4-12 cm broad, dull, continuous; cortex very thin as a polysaccharide 
layer with aculeate hyphae or lacking; algal layer dispersed among superficial medullary 
hyphae; medulla mostly hypophloeodal; apothecia sparse, chroodiscoid, 0°9-1:2 mm diam, 
the thalline rim suberect to recurved, split, exciple free, partially filling the disc; disc white 
pruinose; hymenium c. 130 wm; spores colourless, muriform, 28-30 x 130-150 wm, with 
numerous cells, I—. 


CHEMISTRY. No substances present. 

HaBiTAT. Lower trunks of trees in mossy forest at mid elevations (850-1100 m). 
DISTRIBUTION. Sierra Leone, India, Sri Lanka. 

ADDITIONAL SPECIMENS. Hale collections : 1 (50 198), 5 (46 222, 46 234, 46 251). 


OBSERVATIONS. This member of the ‘7helotrema platycarpum’ group has large spores. It is 
probably close to 7. /leprocarpum Tuck., a New World species, which has somewhat smaller 
spores (less than 100mm) and smaller apothecia. Nylander compared it with T. 
leucophthalmum Nyl., which contains stictic acid, is definitely corticate, and has small 
spores (23-35 um long). 

6. Thelotrema dilatatum (Mill. Arg.) Hale in Mycotaxon 11 : 131 (1980). 

(Fig. 7h) 
Ocellularia dilatata Mill Arg. in J. Linn. Soc. (Bot.) 30: 452 (1895). Type: Brazil, Rio de Janeiro, 


Glaziou 5531 pp. (G—lectotype; BM—isolectotype). 
IconEs. Hale, 1974a : 21 fig. 11a (lectotype) and fig. | 1b—Hale, 1978a: 18 fig. Sh. 


Thallus whitish ashy, about 2 cm broad; cortex lacking; algae scattered among patches of 
medullary hyphae; medulla mostly hypophloeodal; apothecia rather rare, chroodiscoid, 
about | mm diam, the thalline rim coarse, split and recurved, the inner surface and disc 
densely white pruinose, the exciple fused; disc open, white pruinose; hymenium 130 um; 
spores colourless, transversely septate, 6-8 x 30 um, 5-7 loculate, I+. 


CHEMISTRY. Stictic and constictic acids. 

HaBiTAT. Canopy branch in rain forest at low elevations (150 m). 
DISTRIBUTION. West Indies, Central and South America, Sri Lanka. 
ADDITIONAL SPECIMENS. Hale collections: 9 (46 218). 


OBSERVATIONS. This species was previously known only from the New World. It is 
characterized by the lack of a cortex and large, often irregularly elongate apothecia with a 
thick thalline rim. Similar species in Sri Lanka include T. magnificum and T. waasii, which 
have pigmented discs, and T. patens. 


7. Thelotrema dissultum Hale sp. nov. 


(Fig. 7g) 
Thallus corticola, epiphloeodes, obscure viridis, 4-8 cm latus; apothecia conspicua, chroodiscoidea, 
thallum superantia, I-1°5 mm diametro, margine thallino erecto vel recurvato, late fisso, 


subgranulato, excipulo crasso, periphysoidibus instructis, disco albo pruinoso; hymenium 70-90 um 
altum; sporae incolores, murales, octonae, 5-8 x 14-16 wm, 1-2 x 6-7 loculatae, I—. 

Typus: Sri Lanka, Sabaragamuwa Province, Ratnapura District, Halwathura-Kanda, near 
Rassagala, elev. 900 m, 15 Mar. 1978, M. E. Hale 51 139 (US—holotypus; BM—isotypus). 


Thallus dark greenish gray, 4-8 cm broad, shiny, continuous; cortex dense, 15-18 um; algal 
layer continuous, 15-25 wm; medulla to 150 ~m with periderm and crystalline inclusions; 
apothecia conspicuous, 1-1°5 mm diam, the thalline rim erect to barely recurved, thick, 
coarsely split and subgranular at maturity, white, the exciple free, coarse; disc partially 


THELOTREMATACEAE IN SRI LANKA 255 


closed, white pruinose; hymenium 79-90 wm; spores colourless, muriform, 5-8 x 14-16 um, 
1-2 x 6-7 loculatae, I—. 


CHEMISTRY. Stictic acid. 


HABITAT. Trees along trails in mossy forest at mid elevations (900 m). 
DISTRIBUTION. Sri Lanka. 
ADDITIONAL SPECIMENS. Hale collections: 2 (51 213,51 235). 


OBSERVATIONS. The dark waxy thallus provides a sharp backdrop for the large, white 
erupting apothecia. The thalline wall is generally erect with some tendency to recurve. 
There are no close relatives in the genus. 


8. Thelotrema imperfectum Hale sp. nov. 
(Fig. 7h) 


Thallus corticola, epiphloeodes, tenuis, nitidus, pallide brunneus, 8—14 cm latus; apothecia dispersa, 
inconspicua, semi-emergentia vel immersa, 0°5-O°8 mm diametro, margine thallino integro vel 
minute lacerato, decolorato, excipulo distincto, libero, periphysoidibus instructis; ostiolum 
rotundatum, 0°2-0°4 mm diametro; hymenium 110-180 wm altum; sporae incolores, transversim 
septatae, 4-8 : nae, 6-12 x 35-70 wm, 12-18 loculatae, I+ coeruleae. 

Typus: Sri Lanka, Sabaragamuwa Province, Ratnapura District, Sinharaja Forest Reserve near 
Weddagala, elev. 300 m, 14 Mar. 1978, M. E. Hale 50 443 (US—holotypus; BM—isotypus). 


Thallus greenish to olive whitish gray, 8-14 cm broad, thin, continuous; cortex loosely 
organized, 15-20 um; algal layer 10 wm; medulla about 15 wm with numerous crystals; 
apothecia numerous, nearly flush to semi-emergent, 0°5-O°8 mm diam, the thalline rim 
ragged, crumbling apically, suberect, exciple apically free, rather coarse and filling the disc, 
in part as an inner pore, periphysoids present; pore gaping, 0°2-0'4 mm diam; hymenium 
110-180 wm; spores colourless, transversely septate, 6-12 x 35-70 wm, 4—8/ascus, 12-18 
loculate, I+. 


CHEMISTRY. No substances present. 


HaBITAT. Tree base, lower trunk, and saplings in rain forest at low to mid or higher 
elevations (150-2000 m). 


DISTRIBUTION. Sri Lanka. 


ADDITIONAL SPECIMENS. Hale collections: | (50, 193, 50 199, 50 229), 2 (51 172, 51 218, 51 
221), 3a (46 346, 46 362), 4 (47 067, 47 110, 47 125, 47 128, 47 136, 50 378, 50 385, 50 394, 
50 400, 50 409, 50 410, 50 420, 50 422, 50 434, 50 441, 50 489, 50 491, 50 495, 50 518), 5 
(46 233, 46 239, 46 272, 46 276), 7b (50 295, 51 031, 51 085), 9 (46 155), 10 (50 281). 


OBSERVATIONS. It is remarkable that a species as common as this was not collected by 
Thwaites. At first it seemed to be an anomalous population of T. subtile Tuck. but the spores 
are significantly, and consistently, larger (7. subtile has spores 25-35 wm long), the pore is 
ragged and coarsely pruinose, and the apothecia are in general somewhat raised. Another 
species here, T. kamatii (see below), has even larger spores, 120-150 um long, more or less 
flush apothecia, and an entire, barely raised rim, and a fourth one, T. nureliyum (below), 
occurring at higher elevations, has large spores and a coarser exciple. All of these species are 
related by the presence of periphysoids and lack of chemistry. 


_ 9. Thelotrema kamatii (Patw. & C. Kulk.) Hale in Mycotaxon 11 : 131 (1980). 
(Fig. 71) 
Ocellularia kamatii Patw. & C. Kulk. in Norw. J. Bot. 24: 130 (1977). Type: India, Karnataka, 


Carmudi Ghat, M. B. Nagarkar & K. D. Gole 761061 (AMH—holotype; US—isotype). 
IcONES. Patwardhan & Kulkarni, 1977 : 129 fig. 6. 


256 M. E. HALE 


Thallus pale olive greenish, 6-10 cm broad, thin, dull; cortex dense, variable, 10-20 um; 
algal layer 10-15 wm, discontinuous among subcortical crystals; medulla 30-90 um with 
dense crystals; apothecia numerous, round, 0°3-0'4 mm diam, more or less flush to slightly 
raised, the thalline rim entire to crenate, exciple free, more or less filling the pore at 
maturity; pore gaping, c. 0°3 mm diam; hymenium 180-200 um; spores colourless, trans- 
versely septate, 10-25 x 120-150 wm, 16-22 loculate, 2+4/ascus, I+. 


CHEMISTRY. No substances present. 

HaBITAT. Canopy branches in rain forest at low elevations (300 m). 
DISTRIBUTION. India, Sri Lanka. 

ADDITIONAL SPECIMENS. Hale collections: 7a (50 098, 50 166). 


OBSERVATIONS. The relation of this species to T. imperfectum, T. nureliyum, and T. subtile is 
discussed above under 7. imperfectum. The spores of the type collection do react I+ blue, 
although the original description says iodine negative (Patwardhan & Kulkarni, 1977). This 
is one of the few species collected so far only in India and Sri Lanka. 


10. Thelotrema lacteum Krempelh. in Nyl. in Flora, Jena 47 : 269 (1864). 
(Fig. 8a) 


Type: Australia, comm. Hochstetter (M—lectotype). 

Thelotrema cavatum var. dolichosporum Nyl. in Annls Sci. nat. (Bot.) IV, 11: 242 (1859). Type: 
Polynesia, Marquesas, Nuku Hiva, Mercier s.n. (PC—lectotype). 

Thelotrema dolichosporum (Nyl.) Nyl. in Bull Soc. linn. Normandie Il, 2 : 72 (1868). 

Ocellularia dolichospora (Nyl.) Mill Arg. in Mem. Soc. Phys. Hist. nat. Genéve 29(8) : 9 (1887). 

Phaeotrema lacteum (Krempelh. in Nyl.) Mill. Arg. in Flora, Jena 70 : 398 (1887). 

Ocellularia cricota F. Wilson in Bailey in Queensl. Dept. Agric. Bull. 7:32 (1891). Type: Australia, 
Southport, Wilson s.n. (G—lectotype). 

Phaeotrema cricotum (F. Wilson) Mull. Arg. in Hedwigia 32 :130 (1893). 

Ocellularia japonica Zahlbr. in Annls Mycol. 14: 49 (1916). Type: Japan, Prov. Kotsuke, Mt Akagi, 
Tsunoda s.n. (W—lectotype). 

For additional synonymy see Salisbury, 1972a: 270 (as Thelotrema lepadodes) and Hale, 1974a : 29 
(as Phaeotrema disciforme). 

IconES. Salisbury, 19715 : 272 fig. 1 (as Thelotrema lepadodes) and 1972a : 270 fig. 8 (as Thelotrema 
lepadodes).—Hale, 1974a : 30 fig. 14a (as Phaeotrema disciforme). 


Thallus straw coloured, 6-8 cm broad, thin, shiny, rugulose; cortex lacking; algae scattered 
among superficial medullary hyphae; medulla mostly hypophloeodal; apothecia emergent, 
0:7-1:0 mm diam, the exciple free, colourless, pulling away from the wall with the disc; pore 
gaping, 0°2-0°-4 mm diam; hymenium 90-110 4m; spores brown, transversely septate, 
6-10 x 33-60 wm, 7-11 loculate, I—. 


CHEMISTRY. No substances present. 
HABITAT. Lower trunks of trees in rain forest at lower elevations (150-300 m). 


DISTRIBUTION. United States, West Indies, Central America, Kenya, Mozambique, South 
Africa, Assam, India, Sri Lanka, Japan, New Guinea, Australia, Solomon Islands, New 
Caledonia, Polynesia, Hawaii. 


ADDITIONAL SPECIMENS. Hale collections: 7a (50 161), 9 (46 268). Thwaites collections: C. L. 
128 (BM, PDA) (as ‘Lgt. 93. Thelotrema disciforme’) (other duplicates in BM and those in S, 
UPS are Thelotrema monosporum). 


OBSERVATIONS. As one may deduce from the extensive synonymy, this pantropical species 
has not been well understood. It is more or less identical with 7. monosporum (below) except 
for the transversely septate spores. They often occur together in low elevation secondary 
forests. The spores of T. lacteum appear to be smaller, often not more than 65 wm, than those 


THELOTREMATACEAE IN SRI LANKA Dy 


Fig. 8 Species of Thelotrema. (a) T. lacteum (Patwardhan 74:2102 in US). (b) 7. lepadinum (Hale 
50 357). (c) T. leprieurii (Hale 46 123). (d) T. leprocarpoides (Hale 50 294). (e) T. magnificum 
(Hale 46 284). (f) T. monosporum (Patwardhan 73°1477 in US). (g) T. nostalgicum (Hale 50 
343). (h) T. novae-zelandiae (Hale 50 261). (i) T. nurelivum (Hale 50 292). See Fig. 71 for scale. 


258 M. E. HALE 


of 7. monosporum (120-210 um long), but this character needs more study. Salisbury 
(1972a:270) lists Phaeotrema rockii Zahlbr. (= Thelotrema rockii (Zahlbr.) Hale) as a 
synonym but I would prefer to recognize it as a distinct species because it contains stictic 
acid. 


11. Thelotrema lepadinum (Ach.) Ach., Meth. Lich.: 132 (1803). 
(Fig. 8b) 


Lichen lepadinus Ach., Lich. Suec. Prodr.: 30 (1798). Type: Sweden, Acharius s.n. (H-Ach.—lectotype; 
BM, UPS—isolectotypes). 

Thelotrema lepadinum *lepadizum Nyl. in Acta Soc. Scient. fenn. 26:17 (1900). Type: Sri Lanka, 
Peradeniya, A/mquist s.n. (H-Nyl.—lectotype; S—isolectotype). 

For additional synonymy see Salisbury (1972a :267). 

IcONES. Redinger, 1936 : 82 fig. 47.—Salisbury, 1972a : 268 fig. 5.—Hale, 1978a : 40 fig. 12k. 


Thallus whitish to pale tan, 2-6 cm broad, dull, continuous; cortex lacking; algal layer 
scattered among superficial hyphae; medulla mostly hypophloeodal; apothecia emergent, 
0:6-1-2 mm diam, the thalline rim suberect, the exciple free, colourless, partially filling the 
disc; pore gaping, round, 0°2-0°-4 mm diam; hymenium 120-190 um; spores colourless, 
muriform, 15-20 x 65-80 wm, 1-2 x 12-14 loculate, 2-4/ascus, I—. 


CHEMISTRY. No substances present. 
HaBITAT. Lower trunk and branches of trees at higher elevations (1500-2200 m). 


DISTRIBUTION. Canada, United States, Central and South America, Europe, Africa, India, Sri 
Lanka, Japan, Australia, New Zealand, Hawaii. 


ADDITIONAL SPECIMENS. Hale collections: 12 (50 357). Other collections: Nuwara Eliya 
District, Santesson 26 012 (S, US), Moberg 2537 (UPS). 


OBSERVATIONS. This pantemperate weed is not nearly as common in Sri Lanka as it is in the 
nearby mountains of southern India. 


12. Thelotrema leprieurii (Mont.) Hale in Mycotaxon 11 : 131 (1980). 
(Fig. 8c) 


Stictis leprieurii Mont. in Annls Sci. nat. (Bot.) IV, 3 : 97 (1855). Type: French Guiana, Leprieur 804 
(PC—lectotype; G—isolectotype). 

Graphis subnivescens Nyl. in Flora, Jena 69 : 174 (1886). Type: ‘Insulae Guineenses’, Sdo Tomé, 
Moller s.n. (H-Nyl. 7507—lectotype). 

Graphis phlyctidea Vainio in Suomal. Tiedeakat. Toim. A. 15(6): 237 (1921). Type: Philippines, 
Luzon, Sorsogon, Irosin, E/mer 14 646 (TUR—lectotype).. 

Phaeographis phlyctidea (Vainio) Zahlbr., Cat. Lich. Univers. 2 : 383 (1923). 

Phaeographis subnivescens (Nyl.) Zahlbr., Cat. Lich. Univers. 2 : 387 (1923). 

Phaeotrema leprieurii (Mont.) Sherwood in Mycotaxon 5 : 203 (1977). 

For additional synonymy see Salisbury, 19726 : 288 (as Thelotrema leucastrum). 

ICONES. Salisbury, 1972b : 287 fig. 9 (as Thelotrema leucastrum). 


Thallus dark olive green, 5-8 cm broad, shiny, continuous; cortex poorly developed to 
lacking; algal layer 10-15 wm, scattered in the upper medullary layer between crystals; 
medulla 10-30 wm, with numerous periderm and crystalline inclusions; apothecia dispersed, 
chroodiscoid, round to irregularly elongate, 1-2 mm wide, the thalline rim suberect, inner 
wall and tips coarsely white-pruinose, the exciple free at the tips; hymenium 90-110 wm; 
spores brown, transversely septate, 5-10 x 12-18 wm, 4—5 loculate, I—. 


CHEMISTRY. No substances present. 
HABITAT. Open base of trees to canopy in rain forest at lower elevations (150-350 m). 


DISTRIBUTION. West Indies, South America, west Africa, India, Sri Lanka, Sumatra, 
Philippines. 


THELOTREMATACEAE IN SRI LANKA 259 


ADDITIONAL SPECIMENS. Hale collections: 4 (50 397), 7a (50 155), 9 (46 123). 


OBSERVATIONS. The main feature of this species is the conspicuous, sometimes flocculant 
exciple, although this varies and is not highly developed in the Sri Lankan material. When 
Vainio described Thelotrema confluens, he actually compared it with Nylander’s Graphis 
subnivescens. 


13. Thelotrema leprocarpoides Hale sp. nov. 
(Fig. 8d) 


Thallus corticola, hypophloeodes, opacus, albo-cinereus, 8—12 cm latus; apothecia rotundata vel pauce 
elongata, 0°4-0°6 mm diametro, margine thallino vix elevato vel suberecto, excipulo libero, crasso, 
periphysoidibus instructis, disco aperto; hymenium 90 yum altum; sporae incolores, murales, 
4-8 : nae, 10-12 x 36-40 wm, |-3 x 8-10 loculatae, I—. 

Typus: Sri Lanka, Sabaragamuwa Province, Ratnapura District, near Pagoda south of Ratnapura, 
12 Feb. 1976, M. E. Hale 46 381 (US—holotypus; BM—isotypus). 


Thallus whitish mineral gray, 8-12 cm broad, thin, dull, continuous; cortex not developed; 
algal layer distinct, 10-15 um, interrupted by crystals; medullary area in part epiphloeodal, 
10-90 um, consisting mostly of oxalate crystals; apothecia round to somewhat elongate, 
0:-4-0'6 mm diam, solitary or aggregated in twos and threes, the thalline rim barely 
developed, low to suberect and splitting, barely recurved, the exciple free, coarse, partially 
filling the disc; hymenium 90 um; spores colourless, muriform, 10-12 x 36-40um, 
1-3 x 8-10 loculate, 4-8/ascus, I—. 


CHEMISTRY. No substances present. 
HABITAT. Cultivated trees at lower elevations (150-200 m). 
DISTRIBUTION. India, Sri Lanka. 


ADDITIONAL SPECIMENS. Hale collections: 9 (Rest House at Hiniduma, 46 380). India: 
Karnataka, Koppa-Sringeri road, Hale 49 949 (US). 


OBSERVATIONS. This represents another addition to the ‘Thelotrema platycarpum’ group. It 
is distinguished by the pruinose disc and exciple as well as lack of any chemistry. 


14. Thelotrema magnificum (Berk. & Broome) Hale in Mycotaxon 11 : 131 (1980). 
(Fig. 8e) 


Platygrapha magnifica Berk. & Broome in J. Linn. Soc. (Bot.) 14: 110 (1875). Type: Sr Lanka, 
| Thwaites 624 (K—lectotype). 

_ Ocellularia magnifica (Berk. & Broome) Sherwood in Mycotaxon 3(2) : 234 (1976). 

IcONES. Sherwood, 1976 : 237 fig. 2. 


Thallus greenish mineral gray, 10-15cm broad, smooth, continuous; cortex dense, 
15-20 um, with aculeate hyphae, splitting somewhat internally; algal layer continuous, 
15-20 um; medulla 6-15 wm, mostly hypophloeodal; apothecia very conspicuous, chro- 
odiscoid, 1-2 mm diam, the thalline rim coarse, strongly emergent and erect to recurved, the 
exciple fused; disc open, broad, flat, orange-red pruinose; hymenium 150-180 um; spores 
_ colourless, transversely septate, 8-10 x 70-90 wm, 18-22 loculate, 4—-6/ascus, I+. 


_ CHEMISTRY. No medullary substances present; epithelial pigment K+ deep purple. 


HaBITAT. Base and lower trunk of trees in rain forest at lower elevations (150-550 m). 
DISTRIBUTION. Sri Lanka. 


ADDITIONAL SPECIMENS. Hale collections: 1 (50 217), 2 (51 111, 51 156), 3a (46 284, 46 348), 
4 (47 090, 47 112, 50 456, 50 481), 7b (50 296, 51 034), 9 (46 149, 46 168). 


OBSERVATIONS. This conspicuous lichen was described as a fungus and therefore escaped the 


260 M. E. HALE 


attention of both Leighton and Zahlbruckner. Sherwood was the first to determine its correct 
status during a revision of the fungal genus Platygrapha. Berkeley and Broome had labelled 
the type ‘Platygrapha lateritia’ but changed the name before publication. It is still known 
only from Sri Lanka where it is common, more than my 13 collections suggest, since I could 
recognize it in the field by the brilliant orange-red discs and therefore only representative 
specimens were collected. 


15. Thelotrema monosporum Ny. in Annls Sci. nat. (Bot.) IV, 15 : 46 (1860). 
(Fig. 8f) 


Type: New Caledonia, Pancher s.n. (H-Nyl. 22709—lectotype). 

Thelotrema disciforme Leighton in Tran. Linn. Soc. Lond. 27 : 170 (1870). Type: Sri Lanka, Thwaites 
C.L. 128 (BM—lectotype; G, H-Nyl. 3851, PC, PDA, S, UPS, W—isolectotypes (as ‘Lgt. 93. 
Thelotrema disciforme’). 

Leptotrema monosporum (Nyl.) Mull. Arg. in Bull. Soc. r. Bot. Belg. 31 : 35 (1892). 

For additional synonymy see Salisbury (1972a : 271) and Hale (1978a : 52) (as L. lepadodes). 

IcongES. Salisbury, 1972a : 272 fig. 10; Hale, 1978a: 51 fig. 14f(as L. lepadodes). 


Thallus tannish gray, 24 cm broad, dull, continuous; cortex lacking; algae scattered among 
superficial medullary hyphae; medulla mostly hypophloeodal; apothecia emergent, 
0:7-1:1 mm diam, the exciple free, pulling away from the thalline wall with the disc; 
hymenium 100-140 um; spores brown, muriform, 7-12 x 45-60 um, 1-3 x 22-24 loculate, 
2-4/ascus, I—. 


CHEMISTRY. No substances present. 


HaBITAT. Typically found in secondary forests. I did not recollect it in Sri Lanka but it can be 
expected to occur in the drier areas of the island. 


DISTRIBUTION. Pantropical. 


OBSERVATIONS. This species forms part of a complex with emergent apothecia, a 
periphysoid-bearing exciple, and a non-corticate thallus, including at the world level, T. 
lepadinum, T. hawaiiense (Hale) Hale, T. lacteum, and T. rockii (Zahlbr.) Hale. To these we 
might add closely related 7. conferendum Hale (Hale, 1975 : 176) and T. porinoides (below), 
both of which have less emergent, nearly flush apothecia. I am adopting Salisbury’s 
(1975 : 59) interpretation of the type of Thelotrema monosporum for the sake of stabilizing 
the nomenclature instead of using 7. lepadodes, although I feel that the Helsinki material 
(there is none in PC) is so fragmentary that the only certain character is the presence of 
brown spores. 


16. Thelotrema nostalgicum G. Salisbury in Lichenologist 5 : 266 (1972). 
(Fig. 8g) 


Type: Sri Lanka, Pidurutalagala, 1879, A/mquist s.n. (H-Nyl. 22 741—lectotype; S—isolectotype). 
ICONES. Salisbury, 1972a : 266 fig. 3. 


Thallus whitish mineral gray, 3-5 cm diam, shiny, verruculose; cortex probably lacking; 
algal layer scattered among superficial medullary hyphae, to 15 wm; medulla largely 
hypophloeodal; apothecia conspicuous, strongly emergent and urceolate, 1 mm high, 
0:7-1:2 mm diam, the amphithecium smooth, corticated, exciple apically free, barely visible 
around the disc at base of apothecial cavity; pore round, 0:1-0°3 mm diam; hymenium 
180-240 um; spores colourless, transversely septate, 25-33 x 180-200 wm, 25-30 loculate, 
1-4/ascus, I+. 


CHEMISTRY. Fumarprotocetraric acid and traces of protocetraric acid. 
HaBITAT. Twigs of exposed shrubs at high elevations (2400 m). 
DISTRIBUTION. Sri Lanka. 


THELOTREMATACEAE IN SRI LANKA 261 
ADDITIONAL SPECIMENS. Hale collections: 12 (50 343, 50 356). 


OBSERVATIONS. Salisbury described his new species from a fragment collected by Almquist. I 
was able to revisit the type locality and make additional collections. Externally it is close to 
T. pidurutalagalum and T. weberi (see below), two other high elevation species in the 
Nuwara Eliya region with urceolate, periphysoid-bearing apothecia. Thelotrema nostalgi- 
cum is distinguished by the unusual chemistry. 


17. Thelotrema novae-zelandiae Szat. in Borbdsia 1 : 56 (1939). 
(Fig. 8h) 
Type: New Zealand, Waikare-Moana, 1932, J. Jablonzky s.n. (BP—lectotype). 


Thallus tannish gray, 2-3 cm diam, dull, continuous; cortex not clearly developed; algal layer 
scattered in superficial medullary hyphae 10-20 um thick; medulla mostly hypophloeodal; 
apothecia chroodiscoid, 0°7-0°9 mm diam, the thalline rim split, incurved at first, later 
becoming recurved, arching over the disc, the exciple free; hymenium 130-140 um; spores 
colourless, muriform, 12-15 x 50-60 wm, with numerous cells, |-2/ascus, I—. 


CHEMISTRY. Stictic and constictic acids. 

HABITAT. Base of trees in open forest at higher elevations (2100 m). 
DISTRIBUTION. Sri Lanka, New Zealand. 

ADDITIONAL SPECIMENS. Hale collections: 10 (50 261). 


OBSERVATIONS. This is the first report of the species since it was published. The Sri Lankan 
material has somewhat smaller apothecia than the type but the same chemistry and 
incurved, Geaster-like rim to the apothecia. 


18. Thelotrema nureliyum Hale sp. nov. 
(Fig. 81) 


Thallus corticola, ephiphloeodes, pallide stramineo-cinerascens, 3—5 cm latus; apothecia immersa vel 
semi-emergentia, 0°-4-0°6 mm diam, margine thallino erecto, excipulo libero, periphysoidibus 
instructis; hymenium 150-160 um altum; sporae incolores, tranversim septatae, 7-15 x 60-120 um, 
4 : nae, 22-24 loculatae, I+ coeruleae. 

Typus: Sri Lanka, Central Province, Nuwara Eliya District, between Nuwara Eliya and Hakgala, 
elev. 2100 m, 21 Mar. 1978, M. E. Hale 50 292 (US—holotypus). 


Thallus pale yellowish or whitish gray, 3-5 cm broad, appearing rather thick, dull, fissured; 
cortex weakly organized, irregularly pored, 10-15 wm; algal layer 10-12 wm; medulla 10 um 
or more but largely hypophloeodal, with large crystals; apothecia numerous, partially 
immersed in the periderm, 0°4-0°6 mm diam, the thalline rim raised but not erect, exciple 
free, coarse, sometimes partially filling the pore, periphysoids present; pore round, 
0:1-0'2 mm diam, the disc visible inside; hymenium 150-160 um; spores colourless, 
transversely septate, 7-15 x 60-120 wm, 20-24 loculate, 4/ascus, I+. 


CHEMISTRY. No substances present. 

HaBITAT. Tree branches in rain forest at high elevations (850-2100 m). 
DISTRIBUTION. Sri Lanka. 

ADDITIONAL SPECIMENS. Hale collections: 2 (51 208), 10 (50 268). 


OBSERVATIONS. Thelotrema nureliyum is a member of the 7. subtile complex, a difficult 
group of species with periphysoids but no lichen substances. It has large spores, in the range 
of T. kamatii (see above), which has a thicker, pruinose exciple and rather sparse apothecia. 
The few collections available indicate a high elevation habitat for 7. nureliyum in contrast to 
that of 7. imperfectum, a similar but smaller spored species which grows at lower elevations. 


262 M. E. HALE 


19. Thelotrema patens Nyl. in Acta Soc. Scient. fenn. 26 : 17 (1900). 
(Fig. 9a) 


Type: Sri Lanka, Pidurutalagala, Dec. 1879, Almquists.n. (H-Nyl. 22551—lectotype; S—isolectotype). 


Thallus tannish white, 2-6 cm broad, dull, smooth and continuous; cortex lacking; algal 
layer more or less continuous among superficial hyphae, to 15 um; medulla mostly 
hypophloeodal; apothecia dispersed, solitary, chroodiscoid, 1-2 mm diam, the thalline rim 
split and recurved, the inner surface pruinose and subgranular, the exciple fused; disc plane, 
brownish to white pruinose; hymenium 110-180 um; spores colourless, muriform, 
18-25 x 80-125 wm, with numerous cells, 1—-4/ascus, I—. 


CHEMISTRY. Stictic and constictic acids. 


HapsitTAT. On trunks and branches of trees in rain forest at lower mid to high elevations 
(300-2200 m). 


DISTRIBUTION. Sri Lanka, Sarawak. 
ADDITIONAL SPECIMENS. Hale collections: | (50 254), 7b (50 312). 


OBSERVATIONS. I did not recollect this species on Pidurutalagala but found it in lower 
elevation rain forest. It externally resembles T. dilatatum, which has smaller, I+ blue spores. 


20. Thelotrema phlyctidiodes (Mill. Arg.) Hale in Mycotaxon 11 : 132 (1980). 
(Fig. 9b) 


Ocellularia phlyctidioides Mill Arg. in Hedwigia 32 : 130 (1893). Type: Australia, Brisbane, Bailey 
354 (G—lectotype). 

Ocellularia conglomerata Hale in Smithson. Contr. Bot. 16 : 19 (1974). Type: Dominica, Hale 37 959 
(US—holotype). 

Icongs. Hale, 1974a: 18 fig. 9h.—Hale 1978a: 18 fig. Sc. 

Thallus white, 2-8 cm broad, dull and finely granular; cortex lacking; algae scattered among 

superficial hyphae or subcortical; medulla mostly hypophloeodal; apothecia scattered, 

solitary or clumped in twos and threes, flush to semi-emergent, 0°'4-0°6 mm diam, the 

thalline rim inconspicuous, entire to jagged, suberect, the exciple free apically, nearly filling 

the disc at maturity; disc pale tan to white pruinose; hymenium 40-90 wm; spores colourless, 

transversely septate, 4-5 x 12-20 wm, 5-8 loculate, I+. 


CHEMISTRY. Stictic and constictic acids. 
HABITAT. Saplings and lower trunks in rain forest at mid elevations (350-850 m). 
DISTRIBUTION. West Indies, Central America, Sri lanka, Australia. 


ADDITIONAL SPECIMENS. Hale collections: 1 (50 215A), 2 (51 226), 5 (46 250, 46 270, 46 
278), 7b (50 306, 51 051,51 057, 51 094), 8 (47 039, 47 048). 


OBSERVATIONS. While the thalline rim is not reflexed, the disc is sufficiently open to justify 
placing the species in the ‘7. platycarpum’ group. The free exciple fills the disc partially to 
completely, as noted by Miiller Argoviensis. It is certainly related to T. porinoides, which 
has an incurved exciple and thalline rim with much larger spores. Both lack a cortex and 
contain stictic acid, and both are pantropical. 


21. Thelotrema pidurutalagalum Hale sp. nov. 
(Fig. 9c) 


Thallus corticola, pro parte epiphloeodes, tenuis, rimosus, 2-3 cm latus; apothecia conspicua, 
urceolata, 0°6-0'8 mm diametro, excipulo libero, periphysoidibus instructis; ostiolum rotundatum, 
0°15-0°3 mm diametro; hymenium 180-260 um altum; sporae fuscae, transversim septatae, 
12-21 x 90-100 wm, 15-18 loculatae, raro apice longitudinale 1-2 loculatae, 1-2 : nae, I—. 


. | 


THELOTREMATACEAE IN SRI LANKA 


= y i Se RR Pe ame ee 
(b) T. phlyctidioides (Hale 51 051). (c) T. 
pidurutalagalum (Hale 50 347). (d) T. piluliferum (Hale 50 340). (e) T. platycarpoides (Thwaites 
634 in BM). (f) T. platycarpum (Thwaites C. L. 5 in BM). (g) T. platysporum (Hale 46 161). (h) T. 
platysporum (Hale 46 131). (i) T. porinoides (Hale 50 176). See Fig. 71 for scale. 


264 M. E. HALE 


Typus: Sri Lanka, Central Province, Nuwara Eliya District, Pidurutalagala, elev. 2520 m, 21 Mar. 
1978, M. E. Hale 50 347 (US—holotypus). 


Thallus whitish gray, 2-3 cm broad, dull, thin, fissured; cortex not clearly developed and 
probably lacking; algal layer more or less continuous among medullary hyphae; medulla 
15-20 um, with many crystals, mostly hypophloeodal; apothecia conspicuous, emergent and 
urceolate at maturity, 0°6-0°8 mm diam, the exciple free, pulling away from the wall with 
the disc, periphysoids present; pore gaping, 0°15-0°-3 mm diam, white mmmed and 
eventually pruinose or decomposing; hymenium 180-260 um; spores brown, transversely 
septate, 12-21 x 90-100 wm, 15-18 loculate, the ends occasionally 1-2 loculate longitudin- 
ally at maturity, 1-2/ascus, I—. 


CHEMISTRY. No substances present. 

HasiTaT. On twigs of shrubs at high elevations (2520 m). 
DISTRIBUTION. Sri Lanka. 

ADDITIONAL SPECIMENS. Hale collections: 12 (50 352). 


OBSERVATIONS. This species is closely related to Thelotrema nostalgicum (above) and T. 
weberi (below) in apothecial size and emergence as well as presence of periphysoids. It is 
the only species in the group lacking chemistry. 


22. Thelotrema piluliferum Tuck. in Proc. Am. Acad. Arts Sci. 7 : 227 (1868). 
(Fig. 9d) 


Type: Hawaii, Oahu, Waialua Mtns, Mann s.n. (FH-Tuck.—lectotype; G, W—isolectotypes). 
Thelotrema vernicosum Zahlbr. in Annls mycol. 10 : 370 (1912). Type: Hawaii, Koolau Mountains, 
Rock 101 (W—lectotype; FH—isolectotype) 
Thelotrema gibbosum Magnusson in Magnusson & Zahlbr. in Ark. Bot. 31A(1): 53 (1944). Type: 
Hawaii, Maui, Haelaau, | Aug. 1938, Selling 5836 (UPS—lectotype; S—isolectotype). 


Thallus light mineral gray, c. 8 cm broad, shiny, fissured; cortex dense, irregularly thickened, 
5-10 um, with some internal splitting; algal layer continuous, 15-20 um; medulla 
50-100 um with crystals; apothecia numerous, emergent, rarely urceolate at maturity, 
0-7-1:0 mm diam, the exciple fused, yellowish brown; pore tiny, c. 0°05 mm diam, often 
nearly closed, depressed; hymenium 140-160 wm; spores colourless, muriform, 8-12 x 24—- 
28 um, l-3 x 4-6 loculate, uniseriate, I+. 


CHEMISTRY. Psoromic acid. 

HaBITAT. Trees along trail at higher elevations (2200 m). 
DISTRIBUTION. Sri Lanka, Sabah, Hawaii. 

ADDITIONAL SPECIMENS. Hale collections: 12 (50 340). 


OBSERVATIONS. The strongly emergent, tiny-pored apothecia resemble those of Pertusaria, 
as Tuckerman noted in his original description. Salisbury (1972a : 273) presumed it to be 
near 7. lepadinum because of the periphysoids, as reported by Magnusson. As a matter of 
fact, Magnusson correctly identified 7. piluliferum from Hawaii and still described T. 
gibbosum since it ‘belonged to quite another group than the other Hawaiian species’ because 
of the periphysoids. He evidently overlooked the periphysoids in Tuckerman’s type. 

Thelotrema piluliferum as delimited here has a fused exciple with dense short 
periphysoids, but differs considerably from the ‘typical’ species in the genus in having 
psoromic acid and a small pore, both characters more typical of Myriotrema, where one 
might be tempted to place it. Thelotrema platysporum (below) and T. saxicola (Vainio) G. 
Salisb. (Salisbury, 1972a: 269) share these anomalous characters. Three additional rare 
species have periphysoids and psoromic acid: poorly known T. dislaceratum Krempelh. 
from south-east Asia, T. cryptotrema Nyl. and T. sphinctrinellum Nyl. from tropical 
America. 


THELOTREMATACEAE IN SRI LANKA 265 


23. Thelotrema platycarpoides Tuck. in Proc. Am. Acad. Arts Sci. 6 : 270 (1864). 
(Fig. 9e) 


Type: Cuba, 17 Jan. 1863, Wright 157 (FH-Tuck.—lectotype; BM, H-Nyl. 22671, G, L, PC, UPS, 
US, W—isolectotypes). 

Platygrapha bivela Berk. & Broome in J. Linn. Soc. (Bot.) 14 : 109 (1875). Type: Sri Lanka, Thwaites 
634 (K—lectotype; BM—isolectotype). 

Phaeotrema platycarpoides (Tuck.) Miill. Arg. in Flora, Jena 69 : 311 (1886). 

Icongs. Salisbury, 1972 : 287 fig. 8. 


Thallus pale tannish, 6-7cm broad, thin, shiny; cortex dense, 15-18 wm, splitting 
internally; algal layer continuous, 15 wm; medulla 10-30 wm, with crystals; apothecia 
numerous, chroodiscoid, 0°5-1:2 mm diam, solitary or clustered in twos and threes, 
the thalline rim splitting and recurved, exciple free; disc open, rusty brown or becoming 
white pruinose at maturity; hymenium 60-70 um; spores brown, transversely septate, 
5-6 x 11-15 wm, 3-4 loculate, I-. 


CHEMISTRY. Stictic and constictic acids. 
HaBiTAT. Secondary forests at low elevations. 
DISTRIBUTION. Mexico, West Indies, Sri Lanka, Java. 


OBSERVATIONS. This species is well characterized by the chroodiscoid apothecia and 
chemistry. Tuckerman originally separated it from 7. platycarpum because of the smaller 
apothecia (0°5-1‘'2 mm vs 1:0-2'5mm diam) and somewhat larger spores. My own 
measurements on the types and other specimens, however, do not show any significant 
differences in spore size. Leighton added a note to the Thwaites collection of Platygrapha 
bivela in BM to the effect that it was a synonym of ‘ Thelotrema platycarpoides Wright’s Lich. 
Cubae 157’ but the species was not included in his 1870 publication. 


24. Thelotrema platycarpum Tuck. in Proc. Am. Acad. Arts Sci. 5 : 406 (1862). 
(Fig. 9f) 


Type: Cuba, Wright 139 (FH-Tuck.—lectotype; BM, H-Nyl. 22668, G, L, UPS, US— 
isolectotypes). 

Asteristion erumpens Leighton in Trans. Linn. Soc. Lond. 27 : 163 (1870). Type: Sri Lanka, Central 
Province, Thwaites C. L. 5 (BM—lectotype; H-Nyl. 22683, S, UPS—isolectotypes) (as “Lgt. 22, 
Asteristion erumpens’). 

— Platygrapha albo-rufa Berk. & Broome in J. Linn. Soc. (Bot.) 14: 110 (1875). Type: Sri Lanka, 

: Thwaites 69 (K—lectotype). 

Phaeotrema platycarpum (Tuck.) Zahlbr., Cat. Lich. Univers. 2 : 609 (1923). 

Phaeotrema erumpens (Leighton) Santesson in Symb. bot. upsal. 12(1) : 423 (1952). 

Phaeotrema apertum C. Dodge in Beth. nov. Hedwigia 12 : 98 (1964). Type: Uganda, Mulange Forest, 
Dtimmer 4293 (BM—holotype). 

Iconss. Salisbury, 19726 : 286 fig. 7. 


Thallus dark tannish, c. 4cm diam, dull; cortex dense, 15-18 wm, splitting somewhat 
internally; algal layer continuous, to 30 wm; medulla to 30 um, with some crystals; apothecia 
conspicuous, solitary or aggregated, 1-2-5 mm diam, chroodiscoid, the thalline margin 
splitting and strongly recurved, the inner wall white pruinose, the exciple free, erect; disc 
variable, tan to rusty brown or heavily white pruinose; hymenium 60-75 wm; spores brown, 
transversely septate, 4-6 x 9-14 um, 4-loculate, often shrivelling at maturity, I—. 


CHEMISTRY. Stictic and constictic acids. 

HABITAT. Secondary forests at lower elevations. 

DIsTRIBUTION. United States, West Indies, Uganda, Sri Lanka, Sarawak, Java. 
OBSERVATIONS. The large chroodiscoid apothecia are very conspicuous. The differences with 


266 M. E. HALE 


T. platycarpoides are given above. I did not recollect either species in the rain forest; they 
occur in semi-arid secondary habitats. 


25. Thelotrema platysporum Harm. in Bull. Séanc. Soc. Sci. Nancy Il, 13 : 41 (1912). 
(Fig. 9g, h) 


Type: New Caledonia, Pionniero 38 (DUKE—lectotype; FH, W—isolectotypes). 

Thelotrema diminitum Hale in Phytologia 27 : 494 (1974). Type: Sarawak, Bako National Park, 10 
Mar. 1965, M. E. Hale 30 536 (US—holotype) (‘platysporum’ unknown present but psoromic acid 
lacking). 

Icones. Hale, 19746 : 500 fig. 11. 


Thallus whitish or greenish ashy, 6-16 cm broad, shiny, smooth, continuous; cortex dense, 
thin, 5-10 um, with aculeate hyphae, splitting and exfoliating; algal layer partially 
interrupted by crystals, 10-25 wm; medulla 10-110 wm with crystals and cortical inclusions; 
apothecia numerous, emergent, 0°3-0°7 mm diam, solitary or grouped in twos and threes, the 
exciple fused, colourless, with short periphysoids; pore very small, 0:03-0:05 (—0°15) mm 
diam, the surrounding area often crumbling, narrowly black rimmed; hymenium 
90-120 um; spores colourless, muriform, 7-12 x 14-30 wm, 1-2 x 4-5 loculate, uniseriate, 
I+. 


Cuemistry. Unidentified substance above atranorin in t.l.c. plates, with or without psoromic 
acid (confirmed by Dr C. F. Culberson). 


HABITAT. Over mosses and bark on canopy branches in rain forest at low elevations (300 m). 
DISTRIBUTION. Sri lanka, Sarawak, New Caledonia, Australia. 


ADDITIONAL SPECIMENS. Hale collections: 1 (50 249), 2 (51 136), 4 (47 053, 47 057, 47 077, 
47 121, 47 130, 47 161, 50 384, 50 446, 50 466, 50 482), 6 (47 208), 7a (50 156, 50 190), 7b 
(50 302, 51 079), 9 (46 131, 46 172). Thwaites collections: C. L. s.n. (PDA), C. L. 31 (BM, 
H-Nyl. 22 556) (as ‘Lgt. 85. T. punctulatum’), C. L. 99 (BM, H-Nyl. 3880, S, UPS) (as ‘Lgt. 
77. T. sphinctrinellum’), C. L. 174 (BM) (mixed with Myriotrema terebratulum) (as ‘Lgt. 87 
T. myriotrema’). 


OBSERVATIONS. Thelotrema platysporum is an extremely variable species characterized by 
the unknown chemistry and the semi-emergent to emergent apothecia with a fused, 
periphysoid-bearing exciple. The pore area varies from entire and distinctly annulate to 
‘decomposed’ as the rim seems to disintegrate and become lined with protruding excipular 
material. Overall the apothecial structure is similar to that of T. piluliferum and these two 
species are probably very closely related. Thelotrema platycarpum is one of the commonest 
members of the genus in Sri Lanka. 

Leighton identified the Thwaites collection C. L. 99 as ‘T. sphinctrinellum Nyl.’ although 
Nylander marked his duplicate ‘false determinato’. The type of this species from Colombia 
(H-Nyl. with isotypes in FH-Tuck. and PC) has 4-loculate spores and contains psoromic 
acid. Re-examination of these collections led me to conclude that Ocellularia lopezii Hale, 
which I described from Venezuela (Hale, 1975: 174), is a synonym of Thelotrema 
sphinctrinellum. 


26. Thelotrema porinoides Mont. & Bosch in Junghuhn, Enum. Pl. Insul. Java 
Sumatra : 151 (1855). 
(Fig. 91) 


Type: Java, Junghuhn 151 (L—lectotype; FH-Tuck., G, PC, W—isolectotypes). 

Thelotrema exanthismocarpum Leighton in Trans. Linn. Soc. Lond. 27 : 169 (1870). Type: Sri lanka, 
Central Province, Thwaites C. L. 97 (BM—lectotype). 

Thelotrema albidiforme Leighton in Trans. Linn. Soc. Lond. 27: 170 (1870). Type: Sri Lanka, 
Thwaites C. L. 19 (BM—lectotype; PDA—isolectotype). 

Ocellularia albidiformis (Leighton) Zahlbr. in Annis mycol. 14 : 50 (1916). 


THELOTREMATACEAE IN SRI LANKA 267 


Ocellularia exanthismocarpa (Leighton) Zahlbr., Cat. Lich. Univers. 2 : 590 (1923). 
For additional synonymy see Hale (1974a : 20). 
Icones. Hale, 1974a : 21 fig. 1 1c (lectotype).—Hale, 1978a: 18 fig. 51.—Salisbury, 1972a : 265 fig. 2. 


Thallus whitish to pale greenish, 2-8 cm broad, dull, continuous; cortex lacking; algal layer 
scattered in patches of superficial medullary hyphae; medulla mostly hypophloeodal but 
often developed near the apothecia; apothecia common, solitary or aggregated in twos or 
threes, more or less immersed to semi-emergent, to 0°7 mm diam, the thalline wall 
sometimes flaring, the exciple free but distinctly incurved to form an inner pore; pore round, 
0:2-0'4 mm diam; hymenium c.180 um; spores colourless, transversely septate, 8-15 x 
60-125 wm, 26-29 loculate, I+. 


CHEMISTRY. Stictic and constictic acids, rarely accompanied by the higher ‘quintaria’ spot 
(Hale, 1974a: 12). 


HaBiTaT. Tree trunks along trails at 150-2000 m elevation. 


DISTRIBUTION. West Indies, Central America, India, Sri Lanka, Philippines, Java, Sabah, 
Solomon Islands, Hawaii. 


ADDITIONAL SPECIMENS. Hale collections: | (50 200, 50 202, 50 213), 2 (51 181), 3a (46 300, 
46 304), 5 (46 229), 7a (SO 176, 50 188), 8 (47 029), 9 (46 198), 10 (50 263). Thwaites 
collections: C. L. 97 (BM, H-Nyl. 3855) (as ‘Lgt. 91, Thelotrema exanthismocarpum’; PDA 
material is Ocellularia stictidea). 


OBSERVATIONS. The double pore, resulting from the distinct, incurved exciple, has been 
noted by the various authors who studied the species under its numerous synonyms. We 
cannot be sure why Leighton described Thelotrema albidiforme since the description is 
virtually identical with that of T. exanthismocarpum. The scant thallus is ‘lutescens’ since 
the underlying bark substrate is olive brown in contrast to the whitish bark (‘albido- 
fuscescens’) on which the type of 7. exanthismocarpum was growing. In any event I proved 
stictic acid in both collections with t.l.c. 

Salisbury (1972a: 265) places Ocellularia obovata (Stirton) Zahlbr. in synonymy under 
Thelotrema porinoides. The type of T. obovatum from New Zealand (BM), however, is 
corticate, has no inner ring formed by the exciple, and lacks lichen substances. 


27. Thelotrema pseudoexanthismocarpum (Patw. & C. Kulk.) Hale in Mycotaxon 11 : 132 
(1980). 
(Fig. 10a) 


Ocellularia pseudoexanthismocarpa Patw. & C. Kulk. in Norw. J. Bot. 24: 130 (1977). Type: India, 
Kerala, Anamalai Hills, M. B. Nagarkar & K. D. Gole 76°308 (AMH—holotype). 
IcONES. Patwardhan & Kulkarni, 1977 : 129 fig. 7. 


Thallus light yellowish brown, 6-10 cm broad, dull; cortex very dense, 20-30 wm; algal layer 
continuous, 15-20 um; medulla to 30 wm with crystal inclusions, mostly hypophloeodal; 
apothecia numerous, conspicuous, solitary to aggregated, |-1:2 mm diam, the thalline rim 
coarse, erect to nearly recurved, apically lacerated, exciple free, coarse, partially filling the 
disc; pore round, c. 0°1 mm diam but usually lost in the mass of excipular material; 
hymenium 190-200 um; spores colourless, transversely septate, 12-20 x 80-150 wm, 12-15 
loculate, 2/ascus, I+. 


CHEMISTRY. No substances present. 
HaBiTaT. Lower bole to canopy trees in rain forest at lower elevations (150-800 m). 
DISTRIBUTION. India, Sri Lanka. 


ADDITIONAL SPECIMENS. Hale collections: | (50 223), 3a (46 338), 4 (47 115, 50 404), 7b (51 
017), 8 (47 040), 9 (46 165). 


268 M. E. HALE 


Fig. 10 Species of Thelotrema. (a) T. pseudoexanthismocarpum (Hale 47 040). (b) T. 
pulvereodiscum (Hale 47 064). (c) T. scabiomarginatum (Hale 46 265). (d) T. subpatens (Hale 46 
208). (e) T. waasii (Hale 51 175). (f) T. weberi (Santesson 25 980 in US). See Fig. 71 for scale. 


OBSERVATIONS. Ocellularia pseudoexanthismocarpa was first described from India and 
represents another rare example of the Sri Lanka—India distribution pattern. In spite of the 
epithet it differs significantly from Thelotrema porinoides in having a distinct cortex and in 
lacking chemical substances, although the presence of periphysoids places them in a close 
ontogenetic relationship. The specimens from Sri Lanka have larger apothecia than the 
Indian type (Fig. 11g) but preserve the same round, distinct pore within the mass of excipular 
material. Externally it 1s very similar to Thelotrema scabiomarginatum (see below). 


28. Thelotrema pulvereodiscum Hale sp. nov. 
(Fig. 10b) 


Thallus corticola, epiphloeodes, albo-cinereus, opacus, aetate rimosus, 8-10 cm latus; apothecia 
dispersa, chroodiscoidea, rotundata vel irregulariter elongata, 0°8-1-1 mm diametro, margine 
thallino erecto, sorediato, excipulo libero, periphysoidibus instructis, disco nigro vel albo-pruinoso; 
hymenium 60-90 wm altum:; sporae fuscae, murales, octonae, 4 x 9-10 wm, 1-2 x 3-4 loculate, I—. 

Typus: Sri Lanka, Sabaragamuwa Province, Ratnapura District, Sinharaja Forest Reserve near 
Weddagala, elev. 300 m, 13 Feb. 1976, M. E. Hale 47 064 (US—holotypus). 


Thallus whitish gray, 8-10 cm broad, thin, dull, fissuring with age; cortex very thin, 5-7 um; 
algal layer continuous, 10-15 wm; medulla 10 wm or less, mostly hypophloeodal; apothecia 
dispersed, chroodiscoid, round to irregularly elongate, 0°8—1-1 mm, thalline rim low, 


| 


THELOTREMATACEAE IN SRI LANKA 269 


suberect, becoming sorediate at tips and on the inner side, exciple free, thick, partially filling 
the disc, periphysoids present; disc blackish or white pruinose; hymenium 60-90 zm; spores 
brown, muriform, 4 x 9-10 um, 1-2 x 34 loculate, shriveling at maturity, I—. 


CHEMISTRY. Stictic and constictic acids. 
HABITAT. Canopy of trees in rain forest at lower elevations (300 m). 
DISTRIBUTION. Sri Lanka. 


OBSERVATIONS. Brown-spored, chroodiscoid species of Thelotrema are very rare; | know of 
only two others, Thelotrema mirabile (Zahlbr.) Hale, a species from Taiwan first described 
under Phaeographina, characterized by a very broad disc (3 mm diam) and larger spores 
(42-46 um long), and T. stellatum (Hale) Hale (see Hale, 1978a : 54), which has very large 
spores (to 120 wm long) and no chemistry. 


29. Thelotrema scabiomarginatum Hale sp. nov. 
(Fig. 10c) 


Thallus corticola, epiphloeodes, viridi-cinerascens, 8-15 cm latus; apothecia immersa, 0°6-1°5 mm 
diametro, margine thallino erecto, excipulo libero, crasso, periphysoidibus instructis, columella 
nulla; hymenium 120-210 4m altum; sporae incolores, murales, 10-20 x 50-180 um, dense 
cellulares, 1-4 : nae, I—. 

Typus: Sri Lanka, Southern Province, Matara District, mossy forest above Enselwatta, elev. 
1100 m, 14 Feb. 1976, M. E. Hale 46 240 (US—holotypus; BM—isotypus). 


Thallus light greenish to tannish gray, 8-15 cm broad, shiny, continuous or fissured; cortex 
dense, 12-20 wm, with aculeate hyphae; algal layer continuous, 10-15 um; medulla mostly 
hypophloeodal; apothecia conspicuous, immersed, chroodiscoid, 0°6—1°5 mm diam, solitary 
or aggregated in twos, the thalline rim becoming erect, apically broadened and crumbling, 
exciple free, partially or completely filling the disc, periphysoids present, pore round, 
0-1 mm diam but often lost in the erupting mass of excipular material; hymenium 
120-210 um; spores colourless, muriform, 10-20 x 50-180 um, becoming densely celled, 
1-4/ascus, I+. 


CHEMISTRY. No substances present. 
HABITAT. Base and lower trunk of trees in rain forest at low to mid elevations (150-850 m). 
DISTRIBUTION. Sri Lanka. 


ADDITIONAL SPECIMENS. Hale collections: 1 (50 240, 50 253), 2 (51 105), 3a (46 293, 46 312), 
4 (47 065, 47 091, 47 162, 50 432, 50 453, 50 475, 50 519), 5 (46 245, 46 265), 9 (46 170, 46 
173, 46 211). 


OBSERVATIONS. This endemic species is very close to 7. pseudoexanthismocarpum in the 
development of large, erumpent apothecia. The spores are often poorly developed, and some 
of my identifications were made on the basis of apothecial structure only. 


30. Thelotrema subpatens Hale sp. nov. 
(Fig. 10d) 


Thallus corticola, pro parte hypophloeodes, continuus, albo-cinereus, 8-12 cm latus; apothecia 
subemersa, 0°5-O°8 mm diametro, margine thallino suberecto, excipulo libero, periphysoidibus 
instructis, disco aperto; hymenium 60 wm latum; sporae incolores, transversim septatae, 5—6 x 24- 
25 um, 7-8 loculatae, I+ coeruleae. 

Typus: Sri Lanka, Southern Province, Galle District, Kanneliya Forest Reserve, elev. 150 m, 16 
Feb 1976, M. E. Hale 46 208 (US—holotypus; BM—isotypus). 


Thallus whitish mineral gray, 8-12 cm broad, shiny, continuous; cortex barely discernible 
and probably not developed; algae scattered among superficial medullary hyphae; medulla 
to 254m in patches, mostly hypophloeodal; apothecia immersed to barely raised, 


270 M. E. HALE 


0:5-0°8 mm diam, the thalline rim erect to barely recurved, the exciple free, partially filling 
the disc, periphysoids present; hymenium 60 wm; spores colourless, transversely septate, 
5-6 x 24-26 wm, 7-8 loculate, I+. 


CHEMISTRY. No substances present. 
HABITAT. Lower trunks in rain forest at low elevations (150 m). 
DISTRIBUTION. Sri Lanka. 


OBSERVATIONS. This species is represented only by the type specimen. There are no close 
relatives, except perhaps 7. pycnophragmium Ny|l., which has larger, muriform spores. 


31. Thelotrema waasii Hale sp. nov. 
(Fig. 10e) 


Thallus corticola, hypophloeodes, opacus, albidus, 2-3 cm latus; apothecia conspicua, chroodiscoidea, 
disco late aperto, 1-1-5 mm diametro, margine thallino valde recurvato, fisso, onmino purpureo, 
excipulo connato, periphysoidibus instructis, hymenium 65-70 um altum; sporae incolores, 
transversim septatae, octonae, 5 x 15-18 wm, 5-6 loculatae, I—. 

Typus: Sri Lanka, Sabaragamuwa Province, Ratnapura District, Halwathura-Kanda, near 
Rassagala, elev. 850-900 m, 15 Mar. 1978, M. E. Hale 51 175 (US—holotypus). 


Thallus white, 2-3 cm broad, dull, continuous; cortex lacking; algae scattered among 
superficial medullary hyphae; medulla mostly hypophloeodal; apothecia chroodiscoid, very 
conspicuous, |-1°5 mm diam, the thalline rim strongly emergent, split and recurved, coarse, 
the exciple fused, with short periphysoids; disc and inner thalline wall deep violet purple K+ 
dark purple; hymenium 65-70 um; spores colourless, tranversely septate, 5 x 15-18 wm, 5-7 
loculate, I—. 


CHEMISTRY. No medullary substances present; unidentified purple pigment K+ purple. 
HasirTat. Branch of sapling in rain forest at mid elevation (900 m). 
DISTRIBUTION. Sri Lanka. 


OBSERVATIONS. Thelotrema waasii resembles T. dilatatum (above) in many respects: the 
large apothecia with recurved thalline rim, fused exciple, and lack of cortex. It differs in the 
brilliantly coloured disc, smaller spores, and lack of stictic acid. It is named in honour of 
Shelton Waas, whose logistical help was crucial during my excursions 1n Sri Lanka. 


32. Thelotrema weberi Hale in Phytolgia 27 : 497 (1974). 
(Fig. 10f) 


Type: New Guinea, Papua, Eastern Highlands, | July 1968, Weber & McVean 48 666 
(COLO—holotype; US—isotype). 
ICONES. Hale, 19746 : 501 fig. 18. 
Thallus light brownish ashy, c. 6 cm broad, shiny, continuous; cortex probably lacking or 
remaining as a polysaccharide covering; algal layer 15-30 wm; medulla 60-100 wm, with 
large crystals; apothecia conspicuous, emergent and urceolate, 1-1-5 mm diam and to 2 mm 
high, the exciple free, forming an interior ring; pore gaping, 0°3-0°8 mm diam, white 
rimmed; hymenium 250-290 wm; spores colourless, muriform, 15-35 x 150-190 um, 
densely celled, 1-2/ascus, I+. 


CHEMISTRY. Norstictic-acid. 
HABITAT. On cultivated trees (Cupressus) at high elevations (2100 m). 
DISTRIBUTION. Sri Lanka, Sabah, New Guinea. 


ADDITIONAL SPECIMENS. Nuwara Eliya District, Santesson 25 890 (S), Wheeler 12 376, 12 
379 (US). 


THELOTREMATACEAE IN SRI LANKA DT 


OBSERVATIONS. Other comparable periphysoid-bearing species with large urceolate 
apothecia include 7. nostalgicum and T. pidurutalagalum, all found at higher elevations. 
They form a close phylogenetic group in spite of the spore differences. 


Il. MYRIOTREMA Fée 


Essai Crypt. XLIX (1824). 
Coscinedia Massal. in Atti R. Ist veneto Sci. Ill, 5 : 256 (1860). 


Thallus crustose, epiphloeodal; cortex usually dense with splitting and exfoliation, more 
rarely loosely organized or lacking; medulla well developed; apothecia either immersed in 
the medulla without development of a thalline rim, less commonly immersed in the 
periderm, or raised with a distinct incurved thalline rim and a tiny pore; exciple colourless to 
reddish brown, lacking periphysoids, usually fused; spores colourless or brown, transversely 
septate or muriform, I+ deep blue or rarely I—. 


TYPE SPECIES. Myriotrema olivaceum Fée. 
NUMBER OF SPECIES. About 150. 


OBSERVATIONS. The main nucleus of the genus, comprising 21 of the species in Sri Lanka 
related to M. olivaceum, have apothecia immersed in the medulla, 0:2-0°-4 mm diam, and 
flush with the surface: M. album, M. compunctum, M. decorticatum, M. fissurinum, M. 
frondosum, M. glaucescens, M. hartii, M. microporum, M. minutum, M. multicavum, M. 
nuwarense, M. olivaceum, M. polytretum, M. protoalbum, M. rugiferum, M. santessonii, M. 
subconforme, M. terebrans, M. terebratulum, M. thwaitesii, and M. wightii. Six other species, 
M. albocinctum, M. anamalaiense, M. andamanicum, M. masonhalei, M. mastarion, and 
M. minutulum, are deeply immersed in the periderm. 

The remaining species in Sri Lanka have raised apothecia which externally seem more 
related to Ocellularia than to Myriotema: M. cinereoglaucescens, M. costaricense, M. 
desquamans, M. elachistoteron, M. eminens, M. fluorescens, M. glaucophaenum, M. 
granulosum, M. microstomum, and M. porinaceum. They differ from Ocellularia, of course, 
in lacking a carbonized exciple. Only detailed ontogenetic studies, however, can demonstrate 
eventually whether Myriotrema as presently conceived is a polyphyletic group. 

Salisbury (1978) recognized the difference in emergence by dividing the species that he had 
examined into two groups under Thelotrema sect. Myriotrema, the ‘T. album’ group 
(apothecia immersed) and the ‘7. bahianum’ group (apothecia raised). I have not tried to 
organize the Sri Lankan species along these lines. 


Key to the species 

1 Spores colourless . : ; : 4 ; : : 5 ; : ‘ . ; . : D, 
— Spores brown : , : 5 : : : , s : : : : : ; 32 
2 Spores tranversely septate : : : ; A : : : : : : : : ; 3 
— Sporesmuriform. . Sycrad ele ehh wharves ees 21 
3. Spores large, 90-150 um; medulla pale yellow é : s : ; 17. M. granulosum 

— Spores small, less than 30 um long; medulla white ; : E , : f ‘ ; 4 
4 Apothecia deeply immersed inthe periderm . : 5 
— Apothecia immersed in the medulla with little or no periderm inclusions « or somewhat eed 

5 Medulla p— (no substances present). : : : , : i : “L M. albocinctum 

— Medulla P+ yellowororange . : : oF d : : : : : 6 
6 Medulla P+ yellow (psoromic acid present) : ¥ : : : : 23. M. minutulum 

— Medulla P+ orange (‘thwaitesil’ unknown present). : : : 36. M. thwaitesii 

7  Medulla conspicuous, 150-300 wm thick 5 , : 5 : : : : ; , 8 
— Medullacomparativelythin,10-60uwmthick. . . . . . . .) , 13 
8  Thallus conspicuously fissured (without magnification) ; ' : : ; : : 9 


Thallus continuous, fissured only withage . . : s : ; : : : : 10 


M. E. HALE 
Medulla P+ yellow (psoromic acid) . : ; ! : : ‘ 21. M. microporum 
Medulla P— (‘olivacea’ unknown) . ; 27.M. olivaceum 
Pore very tiny, 0:05 mm diam; medulla P+ red (fumarprotocetraric acid) 24. M. minutum 


Pore about 0°! mm diam; medulla P— or P+ yellow 


Medulla P— (‘olivacea’ unknown) . : ; : : ; : : 27.M. olivaceum 
Medulla P+ yellow (psoromic acid) . P ‘ . : 

Thallus surface grainy, friable; medulla with dense crystal inclusions 25.M. multicavum 
Thallus smooth, continuous, not friable; crystal inclusions not dense 35.M. terebratulum 
Apothecia raised, large, 0°7-1:0 mm diam; pore 0°4-0'7 mm diam . 7.M. costaricense 


Apothecia immersed to flush, less than 0°5 mm diam; pore 0:05-0-1 mm diam 
Medulla P— (no substances present) 
Medulla P+ yellow or orange 


Apothecia somewhat raised, 0°3-0- 5 mm diam a le 5. M. cinereoglaucescens 
Apothecia flush, 0°1-0°-4 mmdiam . : ; ; , f : ; 

Thallus pale whitish green; pore 0°! mm diam : : ; : : Dey ™M. album 
Thallus dark olive green; pore 0°01 mm diam 5 : . ; 30. M. protoalbum 
Stictic acid present on t.l.c. : ee ne 34. M. terebrans 


Psoromic acid or ‘thwaitesi1’ unknown present 
Psoromic acid present on t.l.c. 
“TMNT WOON OREO MEIKC 4 6 6 68 56 6 5» 8 5 oe «= 
Thallus 50-200 um thick; apothecia flush. : 25. M. terebratulum 
Thallus less than 100 wm thick; apothecia somewhat raised at maturity 

16.M. glaucophaenum 
Thallus dull; apothecia partially immersed in periderm : ; 36. M. thwaitesii 
Thallus shiny; apotheciaimmersedinmedullaonly . . . . 28. M. polytretum 
Thallus isidiate to pustulate : 
Thallus smooth to verruculose, lacking isidia and pustules : 


Isidia pustular, erupting apically . : ; é : 32. M. santessonii 
Isidia solid and cylindrical or flattened, not pustular sim | yin ag eee ee 

Isidia cylindrical, coarse. ; ; : : : 5 : ; , : 18. M. hartii 
Isidia flattened, leaf-like .  . ; 14. M. frondosum 
Apothecia elongate at maturity; lichexanthone present (thallus orange under ultraviolet 


light) ; : : : : é , : 13. M. fluorescens 
Apothecia round; lichexanthone lacking . : - : : : : 
Spores large, 100-250 um long . 
Spores small to medium sized, 20-75 wm Jong 
Apothecia distinctly emergent . 
Apothecia immersed and flush to barely raised sols 0 ye (be eee 
Apothecia becoming urceolate, 0°-4—0°5 mm diam . é : : : 11. M. eminens 
Apothecia not becoming urceolate, about | mmdiam . 22. M. microstomum 
Apothecia deeply immersed in periderm, flush; medulla P+ yellow (psoromic acid) 

19. M. masonhalei 
Apothecia immersed in the medulla, becoming somewhat raised at maturity; medulla P+ 


orange (norstictic acid present) . ‘ : fe 29.M. porinaceum 
Apothecia deeply immersed in periderm; spores to Wd) um long : 3. M. anamalaiense 
Apothecia immersed in the medulla; spores less than 20 wm long : : ‘ : 

Thallus with sparse, blunt isidia F ; , : ; : § : . 18. M. hartii 
Isidia lacking 5 2 j : : 3 : ; : ; : 

Medulla P+ yellow (psoromic acid) . slag bet! Gira foe 31. M. rugiferum 
Medulla P—(no substances present) : : ; ; i : 33. M. subconforme 
Spores transversely septate only ; : 3 é : : . ; 20. M. mastarion 


Spores muriform 

Apothecia deeply immersed i in periderm: ‘thallus P+ yellow (psoromic acid) A 
4.M. andamanicum 
Apothecia immersed in the medulla; thallus P+ orange (stictic acid) or P— 

Spores large, 45-90 um 

Spores small, 10- -33 um 

Thallus surface grainy and finely verruculose; pore 0: 05-0: | mm 1 diam 10. M. -elachistoteron 
Thallus surface smooth and continuous; pore 0°! mm diam or more . 26. M. nuwarense 


18 
20 


30 
31 


33 


34 
35 
36 


THELOTREMATACEAE IN SRI LANKA 273 


Some Medullaithick andicolummnar,150=600 um =2 99 os 6 ©. ee oe owes Si 
-  Medulla less than 100 um thick,notcolumnar . ; ee ee 39 
37 Medulla usually containing small masses of red crystals, p— : : 37. M. wightii 
— Méedulla white, lacking red crystals, P+ orange (stictic or norstictic acids present) . ; 38 
38 Apothecia round; norstictic acid present ont.l.c. . ; : L 6. M. compunctum 
—  Apothecia angular; sticticacid present ont.l.c.  . é 3 15. M. glaucescens 


39 Cortex and medulla sometimes lacking away from apothecia; pore area darkened . 
8. M. decorticatum 


— Cortex and medulla distinct, uniformly developed; pore area concolourous or white. 40 
40 Apothecia0°3-0'4 mmdiam,raisedat maturity . . . .. . 9. M. desquamans 
—  Apothecia0:2-0°3 mmdiam, flushat maturity  . : ; : : 12. M. fissurinum 


1. Myriotrema albocinctum Hale sp. nov. 
(Fig. 1 1a) 


Thallus corticola, epiphloeodes, minute verrucosus, nitidus, pallide albo- vel viridi-cinereus, 8-12 cm 
latus; apothecia dispersa, immersa, 0°3-0°4 mm diametro, columella nulla; ostiolum rotundatum, 
0:05-0:1 mm diametro, vix annulatum; hymenium c. 90 wm altum; sporae octonae, 5-6 x 16-30 um, 
5-8 loculatae, I+ coerulaeae. 

Typus: Sri Lanka, Sabaragamuwa Province, Ratnapura District, Gilimale Forest Reserve, elev. 
150 m, 12 Feb. 1976, M. E. Hale 46 331 (US—holotypus) 


Thallus pale greenish to whitish gray, 8-12 cm broad, the surface smooth to finely 
verruculose, continuous; cortex dense, 8-l0 um, with aculeate hyphae; algal layer 
continuous, 104m; medulla 10-20 um; apothecia deeply immersed in the periderm, 
0:3-0'4 mm diam, the exciple fused, apically reddish brown, columella lacking; pore round, 
0:05-0:08 mm diam, white rimmed; hymenium 90-140 um; spores colourless, transversely 
septate, 5-6 x 16-30 wm, 5-8 loculate, I+. 


CHEMISTRY. No substances present. 


HaBITAT. Base and lower level, saplings, rarely into the canopy, in rain forest at lower 
elevations (150-450 m). 


DISTRIBUTION. Sri Lanka. 


ADDITIONAL SPECIMENS. Hale collections: 4 (47 099); 6 (47 180, 47 182), 7a (50 146, 50 168), 
7b (50 298, 50 314, 51 009, 51 242), 8 (47 028), 9 (46 182). 


OBSERVATIONS. The deeply sunken apothecia are characteristic of this species, so far known 
only from Sri Lanka. It is related to the psoromic acid-containing M. minutulum. 


2. Myriotrema album Fée, Ess. Crypt. : 104 (1824). 
(Fig. 1 1b) 


Type: South America (?), on Bonplandia trifoliata (G—lectotype; H-Nyl. 22 635, isolectotype). 
Thelotrema album (Fée) Nyl. in Annls Sci. nat. (Bot.) V,7 : 315 (1867). 
Ocellularia alba (Fée) Mill. Arg. in Mém. Soc. Phys. Hist. nat. Genéve 29(8) : 6 (1887). 
See Hale (1978a : 12) for additional synonymy. 
IcOnES. Redinger (1936 : 44) fig. 20; Salisbury (1978 : 411) fig. 5; Hale (1978a : 3, 13) figs Ic (lectotype) 
and 4b. 


Thallus pale greenish gray, 6-10 cm broad, the surface shiny, smooth, somewhat fissured 
with age; cortex cellular, dense, 10-20 wm, with aculeate hyphae, splitting with some 
exfoliating layers; algal layer continuous, 12 um; medulla to 30 um thick with large 
dispersed crystals, mostly hypophloeodal; apothecia numerous, immersed in the medulla, 
0:2-0'4 mm diam, the exciple mostly fused with only the tips free, pale reddish brown, 
prominent to barely visible within the pore; pore round, 0°1-0°15 mm diam, rarely white 
rimmed and more or less raised; hymenium 70-80 um; spores colourless, transversely 
septate, 4-6 x 9-15 wm, 4-loculate, I+. 


274 M. E. HALE 


Fig. 11 Species of Myriotrema. (a) M. albocinctum (Hale 46 331). (b) M. album (Hale 46 190). (c) 
M. anamalaiense (Hale 46 220). (d) M. andamanicum (Hale 51 016). (e) M. cinereoglaucescens 


(Hale 46 363). (f) M. compunctum (Hale 51 033). (g) M. costaricense (Hale 46 143). (h) M. 
decorticatum (Hale 46 259). (i) M. desquamans (Hale 47 209). See Fig. 71 for scale. 


THELOTREMATACEAE IN SRI LANKA 2S) 


CHEMISTRY. No lichen substances present. 
Hasirat. Mid bole to canopy at lower elevations in rain forest (150-300 m). 


DISTRIBUTION. West Indies, Central and South America, Philippines, Sarawak, Solomon 
Islands, Australia. 


ADDITIONAL SPECIMENS. Hale collections: 4 (47 068, 47 078), 9 (46 171, 46 190). 


OBSERVATIONS. Myriotrema album has a thin thallus, usually a rimless pore, and no lichen 
substances. The other confusable species, which occupy the same habitats and elevational 
range, have thalli 200-300 wm thick and different chemistry: M. microporum and M. 
terebratulum with psoromic acid and M. olivaceum with the ‘olivacea’ unknown. 

In my study of the Thelotremataceae of Panama I considered Thelotrema myrioporum 
Tuck. to be a synonym of Ocellularia (Myriotrema) alba. As Tuckerman and later Redinger 
(1936 : 42) discovered, this species has consistently 2-loculate spores, a fact I have confirmed 
both in the type (FH-Tuck.) and in Redinger’s Malme collections (S). What I have identified 
as M. album always has 4-loculate spores, feeling that septation alone is not a strong 
character since other biloculate species (e.g. M. uniseptatum (Hale) (Hale, 1978a : 33) may 
have occasional 3-loculate spores. It is probably better, however, to recognize M. 
myrioporum as a distinct species. 


3. Myriotrema anamalaiense (Patw. & C. Kulk.) Hale in Mycotaxon 11 : 132 (1980). 
(Fig. 1 1c) 
Thelotrema anamalaiense Patw. & C. Kulk. in Norw. J. Bot. 24: 127 (1977). Type: India, Kerala, 
Anamalai Hills, Patwardhan & Kulkarni 76. 344 (AMH—holotype; US—isotype). 
Thallus pale tannish gray, c. 8 cm broad, smooth, continuous; cortex loosely organized, 
10-18 wm, irregularly pored; algal layer 10-15 um; medulla 30-60 wm with numerous 
crystals; apothecia immersed in the periderm, 0°3-0°5 mm diam, exciple partially free, 
usually visible as an inner ring through the pore; pore round, 0°1-0:2 mm diam, depressed at 
maturity, white rimmed; hymenium 150-170 um; spores colourless, muriform, 15-20 x 70- 
75 um, 2-4 x 12-15 loculate, 2-4/ascus, I+. 


CHEMISTY. Stictic and constictic acids. 

HABITAT. Lower trunks of trees in rain forest at mid elevations (1100 m). 
DISTRIBUTION. India, Sri Lanka. 

ADDITIONAL SPECIMENS. Hale collections: 5 (46 220). 


OBSERVATIONS. This species is distinguished by the immersed apothecia, large spores, and 
chemistry. Externally it resembles MM. andamanicum, which has small brown muriform 
spores and psoromic acid. 


4. Myriotrema andamanicum (Nyl.) Hale in Mycotaxon 11 : 132 (1980). 
(Fig. 11d) 
Thelotrema andamanicum Nyl. in Bull. Soc. linn. Normandie Wl, 7: 167 (1873). Type: India, 


Andaman Islands, 1867, Kurz s.n. (H-Nyl. 22 455—lectotype). 
Leptotrema andamanicum (Nyl.) A. L. Sm. in J. Linn. Soc. Lond. 46 : 74 (1922). 


Thallus pale tannish gray, 6-8 cm broad, shiny, continuous; cortex loosely organized, 
10-15 wm; algal layer continuous, 10-15 ~m; medulla 10-30 wm, with dense crystals, mostly 
hypophloeodal; apothecia deeply immersed in the periderm, 0°8-1:1 mm diam, the exciple 
fused, pale reddish brown; pore round, 0°1-0°2 mm diam, with a raised, whitish rim; 
hymenium 150-190 um; spores brown, muriform, 9-15 x 12-18 wm, 1-2 x 3-5 loculate, I—. 


CHEMISTRY. Psoromic acid or no substances present. 
HasitatT. Lianas and canopy branches in rain forest at lower elevations (150-350 m). 


276 M. E. HALE 


DISTRIBUTION. India, Sri Lanka, Philippines. 


ADDITIONAL SPECIMEN. Hale collections: 5 (46 242) (no substances present), 6 (47 210), 7a 
(50 094), 7b (51 005, 51 016, 51 238), 8 (47 014) (no substances present). 


OBSERVATIONS. The pore is often very deep, much as in M. mastarion, a related species. A. 
L. Smith made the combination in Leptotrema when identifying some collections from New 
Caledonia, but I have not been able to locate the specimen in BM for verification. 


5. Myriotrema cinereoglaucescens (Vainio) Hale in Mycotaxon 11 : 132 (1980). 
(Fig. 1 le) 


Thelotrema cinereoglaucescens Vainio in Suomal. Tiedeakat. Toim. A. 15(6): 189 (1921). Type: 
Philippines, prov. Sorsogon, Irosin, Mt Bulusan, Nov. 1915, E/mer 14957 p.p. (TUR—lectotype; 
BM, FH, L, W—isolectotypes). 

Ocellularia cinereoglaucescens (Vainio) Zahlbr., Cat. Lich. Univers. 2 : 586 (1923). 
Thallus dull greenish gray, c. 10 cm diam, shiny, continuous; cortex dense, 10-20 um, with 
abundant aculeate hyphae, splitting and exfoliating; algal layer continuous, 10-15 wm; 
medulla 20-30 wm with numerous crystals; apothecia abundant, immersed in the medulla, 
0:3-0'5 mm diam, becoming slightly raised, exciple fused, colourless, columella lacking; 
pore round, about 0°1 mm diam; hymenium 45-50 wm; spores colourless, transversely 
septate, 4 x 10 um, poorly developed, 5-6 loculate, I+. 


CHEMISTRY. No substances present. 

HaBiTAT. Lower trunk in rain forest at 100-200 m elevation. 
DISTRIBUTION. Sri Lanka, Philippines, Hawaii. 

ADDITIONAL SPECIMENS. Hale collections: 3a (46 363). 


OBSERVATIONS. This is the first report of the species since the original description. It is 
probably close to M. album but the apothecia are always slightly raised. I did not find a 
columella, even in the type collection, contrary to Vainio’s description. Vainio found spores 
8-9 x 18-22 wm. The Hawaiian specimen (Forbes 2229 in US) has spores 18 wm long. The 
Sri Lankan material is obviously immature. 


6. Myriotrema compunctum (Ach.) Hale in Mycotaxon 11 : 133 (1980). 
(Fig. 11f) 
Urceolaria compuncta Ach., Meth. Lich.: 143 (1803). Type: Indonesia, Amboyna, Christopher Smith 
s.n. (LINN-Sm. 1692:7—lectotype; H-Nyl. 22447—isolectotype). 
Leptotrema compunctum (Ach.) Mill. Arg. in Flora, Jena 71 : 527 (1888). 
For additional synonymy see Salisbury, 19715 : 275. 
ICONES. Salisbury, 19716 : 276 fig. 5 and 1978 : 416 fig. 10. 


Thallus white, brownish, or turning dull red, 4-6 cm broad, dull, continuous; cortex lacking 
but a thin irregularly pored polysaccharide layer often present; medulla compact and 
cellular, to 150 wm, with algae located between vertical arrays, crystals abundant; apothecia 
immersed, 0°15-0°2 mm diam, the exciple apically free, colourless; disc pulling away from 
the wall, dark brown to white pruinose; pore round to irregular, 0:05-0:08 mm diam, 
partially filled with excipular material; hymenium 30-35 wm; spores brown, muriform, 
8 x 10-16 wm, 1-2 x 4-6 loculate, I—. 


CHEMISTRY. Norstictic acid and traces of stictic acid. 
HABITAT. Trees at edge of scrub forest at low elevations. 


DISTRIBUTION. United States, West Indies, Central and South America, Angola, 
Mozambique, Sri Lanka, Java, Philippines, Indonesia. 


THELOTREMATACEAE IN SRI LANKA LIT 


ADDITIONAL SPECIMENS. Buttawa, Ruhuna National Park, Fosberg et al. 51 033 (US). 
Thwaites collections: C. L. s.n. (as ‘Lgt. 85. Thelotrema punctulatum’) (BM, S, UPS). 


OBSERVATIONS. This is one of the best known species in the family, collected widely and 
usually correctly identified. It occurs in semi-arid forest and in disturbed areas which I did 
not visit in Sri Lanka. Leighton sent a specimen (C. L. 31 = Thelotrema platysporum) to 
Nylander, who called it ‘7. punctulatum Nyl. L. Exot. 222.’ The remaining material, which 
Nylander did not see, is M. compunctum. 


7. Myriotrema costaricense (Miill. Arg.) Hale in Mycotaxon 11 : 133 (1980). 
(Fig. 11g) 


Ocellularia costaricensis Mill. Arg. in Bull. Soc. r. Bot. Belg. 30 : 75 (1891). Type: Costa Rica, Baie de 
Sabines, 1890, Pittier s.n. (G—lectotype; US (as 5321), W—isolectotypes). 

Ocellularia alba f. costaricensis (Mill. Arg.) Redinger in Ark. Bot. 28A(8) : 45 (1936). 

Rhabdodiscus costaricensis (Mill. Arg.) Vainio in Suomal. Tiedeakat. Toim. A, 15(6): 184 (1921). 


Thallus pale greenish to whitish gray, 8-15 cm broad, shiny, continuous or cracked at 
maturity; cortex dense, 15 wm thick, with aculeate hyphae, splitting and exfoliating; algal 
layer 10-15 wm; medulla to 100 wm with remnants of cortical layers and numerous crystals; 
apothecia common, semi-emergent, 0°4-0'7 mm diam, the exciple free apically; disc open, 
more or less fissured with superficial columella-like bridging; hymenium 55-60 um; spores 
colourless, transversely septate, 4 x 10-15 wm, 5-6 loculate, I—. 


CHEMISTRY. Psoromic acid. 

HABITAT. Lower to mid trunk in rain forest at low elevations (150-300 m). 
DISTRIBUTION. Costa Rica, Sri lanka. 

ADDITIONAL SPECIMENS. Hale collections: 6 (47 199), 9 (46 137, 46 143). 


OBSERVATIONS. Miiller recognized the essential features of this species: raised apothecia, the 
wide pore with an essentially open disc, and the disc ‘subreticulatis fenestratum’. The 
‘columellate’ structures are not carbonized and for this reason Vainio was in error when he 
transferred it to Rhabdodiscus. The material from Sri Lanka matches the Costa Rican type 
well, but I would hardly claim that the species is well understood. Although Redinger 
recognized it as a variety of Ocellularia (Myriotrema) alba, that species lacks any lichen 
substances and has flush apothecia. Miiller himself placed it near Ocellularia viridi-alba 
(Krempelh.) Miill. Arg. (= Myriotrema viridi-album (Krempelh.) Hale), but the type of 
Thelotrema viridi-album in M has muriform, colourless sproes, lichexanthone, and coarse 
isidia. 


8. Myriotrema decorticatum Hale sp. nov. 
(Fig. 11h) 


Thallus corticola, pro parte hypophloeodes, albidus, nitidus, continuus, 6-10 cm latus; apothecia 
immersa, aetate pauce elevata, 0:2-0-4 mm diametro, excipulo connato apice pallide obfuscato; 
ostiolum rotundatum, nigro-cinctum, 0:05-0:1 mm diametro; hymenium 90-150 wm; sporae fuscae, 
murales, octonae, 6-12 x 18-24 um, 1-3 x 5-7 loculate, I—. 

Type: Sri Lanka, Matara District, Enselwatta, virgin mossy forest, 14 Feb. 1976, M. E. Hale 46 259 
(US—holotypus; AMH, PDA—isotypi). 

Thallus white to ashy gray, thin and in part hypophloeodal, shiny, continuous, 6-10 cm 

broad; cortex very thin, c. 5 um, medulla variable, 10-100 um; apothecia immersed in the 

medulla, slightly raised at maturity, 0°2-0-4 mm diam, often clustered, the exciple fused, 
turning brownish apically; pore round, dark rimmed, 0:05-0-1 mm diam; hymenium 

90-150 um; spores brown, muriform, 6-12 x 18-24 wm, 1-3 x 5-7 loculate, I—. 


CHEMISTRY. Stictic and constictic acid. 


278 M. E. HALE 


HaBitTAaT. Lower bole to canopy of trees in virgin rain forest at 150-1000 m elevation. 
DISTRIBUTION. Sri Lanka. 
ADDITIONAL SPECIMENS. Hale collections: 3a (46 303), 6 (47 204). 


OBSERVATIONS. The poorly developed cortex is a consistent character for this inconspicuous 
species. All other species of Myriotrema that I have seen have a well developed cortex. The 
closest relative is probably M. fissurinum (below), and I had in fact first identified it as that 
species. 


9. Myriotrema desquamans (Mill. Arg.) Hale in Mycotaxon 11 : 133 (1980). 
(Fig. 1 11) 


Anthracothecium desquamans Mill. Arg. in Flora, Jena 71 : 48 (1888). Type: Australia, Johnstowns 
River, North Queensland, 1883, F. v. Miiller s.n. (G—lectotype). 

Thelotrema irosinum Vainio in Suomal. Tiedeakat. Toim. A, 15(6): 174 (1921). Type: Philippines, 
Prov. Sorsogon, Irosin, Oct. 1915, Elmer 14 749 (TUR—lectotype; FH, G, L, W—isolectotypes). 

Leptotrema irosinum (Vainio) Zahlbr., Cat. Lich. Univers. 2 : 635 (1923). 

Leptotrema desquamans (Mill. Arg.) Patw. & Makh. in Bryologist 83 : 368 (1980). 


Thallus pale greenish mineral gray or straw coloured, 6-12 cm broad, shiny, continuous; 
cortex dense, 15-20 um, with short aculeate hyphae, splitting and exfoliating; algal layer 
10-15 wm, interrupted by crystals; medulla 30-50 um, with dense crystals and remnants of 
cortical sheets; apothecia partially immersed in periderm, slightly emergent, concolorous 
with the thallus and inconspicuous, 0°3-0°4 mm diam, the exciple fused, reddish brown; 
pore very tiny, 0°05 mm diam, reddish or white rimmed; hymenium 140-160 wm; spores 
brown, muriform, 10-15 x 20-33 um, 1-3 x 5-7 loculate, I—. 


CHEMISTRY. Stictic and constictic acids. 


HaBiITAT. Lianas, lower trunk, rarely in the canopy of trees in rain forest at low to mid 
elevations (150-850 m). 


DISTRIBUTION. India, Sri Lanka, Philippines, Sabah, Australia. 


ADDITIONAL SPECIMENS. Hale collection: 2 (51 162, 51 166, 51 223), 3a (46 335, 46 342, 46 
349), 6 (47 207, 47 209), 7a (50 151, 50 154, 50 187), 7b (50 304, 51 049, 51 080), 8 (47 003, 
47 006, 47 022, 47 026), 9 (46 163). Thwaites collections: C. I. 144 (BM, PDA, S, UPS), C. L. 
281, 284 (PDA) (as ‘Lgt. 86. Thelotrema phaeosporum’); C. L. 175 (UPS) (as ‘Lgt. 87. 
Thelotrema myriotrema’). 


OBSERVATIONS. The tiny pore misled Miiller to describe this species as an Anthracothecium. 
He did note that the thallus was ‘plaguliformi-desquamans’ and cross sections show con- 
siderable exfoliation. It is a member of a large, stictic acid-containing complex which 
includes M. deceptum (Hale) Hale, M. phaeosporum (Nyl.) Hale, M. reclusum (Krempelh.) 
Hale, and M. trypaneoides (Nyl.) Hale, all with a tiny pore and easily confused with non- 
carbonized pyrenocarpous genera. More distantly related are M. elachistoteron and M. 
fissurinum (see below) and M. subcompunctum (Nyl.) Hale. Nylander identified C. L. 144 as 
‘Thelotrema phaeosporum’, which differs in having immersed, flush apothecia. Myriotrema 
desquamans 1s distingushed by the semi-emergent apothecia and small spores. It is one of the 
most commonly collected species of Myriotrema in Sri Lanka. 


10. Myriotrema elachistoteron (Leighton) Hale in Mycotaxon 11 : 133 (1980). 
(Fig. 12a) 


Thelotrema elachistoteron Leighton in Trans. Linn. Soc. Lond. 27: 169 (1870). Type: Sri Lanka, 
Central Province, Thwaites C. L. 132 (BM—lectotype). 
Leptotrema elachistoteron (Leighton) Patw. & C. Kulk. in Norw. J. Bot. 24 : 128 (1977). 


Thallus pale tannish gray, 3-5 cm broad, grainy to verruculose, continuous; cortex dense, 


THELOTREMATACEAE IN SRI LANKA 279 


10-15 wm, bulging out above large crystals; algal layer 10-15 wm, interrupted by crystals; 
medulla to 30 wm, mostly hypophloeodal; apothecia numerous, immersed in the medulla, 
0:25-0'3 mm diam, slightly raised at maturity, the exciple fused to partially free at the tips, 
colourless; pore 0:05-0:1 mm diam, sometimes darkened; hymenium 150-180 wm; spores 
brown, muriform, 25 x 80-90 um, with numerous locules, 24/ascus, I—. 


CHEMISTRY. Stictic and constictic acids. 

HaBiTaT. Lower trunks in rain forest at lower elevations (300 m). 
DISTRIBUTION. Sri Lanka. 

ADDITIONAL SPECIMENS. Hale collections: 7a (50 170). 


OBSERVATIONS. Although M. elachistoteron was described from Sri Lanka, it is very rare 
there; I collected it only once. Duplicates of the type collection (‘Lgt. 92’) are all mixtures of 
M. fissurinum (see below) and only the small collection in BM can be used for typification. 
Basically the most important characters are the medium to large spores and the non- 
exfoliating cortex with a somewhat grainy or finely verruculose surface. I believe that 
Leptotrema microglaenoides (Vainio) Zahlbr. is a synonym (see Hale, 1974a : 40) but would 
like to see more specimens from the New World. In my discussion of that species in 
Dominica I mentioned that Leptotrema (Thelotrema) monosporum had stictic acid. This 
determination was based on a specimen identified by Nylander in PC, but it is not part of the 
type collection. Thelotrema monosporum (see above) lacks lichen substances. Salisbury 
(1975 : 61) adds Leptotrema subgeminum (Nyl.) Zahlbr. as a synonym but I would hesitate 
to confirm this on the basis of the very poor fragment in H-Nyl., even though the chemistry 
and spores are similar. 


11. Myriotrema eminens (Hale) Hale in Mycotaxon 11 : 133 (1980). 
(Fig. 12b) 

Thelotrema eminens Hale in Mycotaxon 3: 177 (1975). Type: Malaya, Pahang, Gunong Brinchang, 3 

Mar. 1965, M. E. Hale 29 947 (US—holotype). 
Icongs. Hale, 1975 : 181 fig. 9. 
Thallus pale ashy green, 6-10 cm broad, shiny, fissured, sometimes coarsely isidate; cortex 
very thin, 5 wm, with aculeate hyphae; algal layer continuous, 10-15 um; medulla 30-60 wm 
with numerous crystals; apothecia clumped, emergent to urceolate, 0-4-0°5 mm diam, the 
exciple free apically, reddish brown, often visible through and partially filling the pore; pore 
round, 0°1-0°'15mm diam; hymenium 190-200 um; spores colourless, muriform, 
20-30 x 120-150 wm, with numerous locules, 2—-4/ascus, I+. 


CHEMISTRY. Stictic acid, probably with traces of norstictic acid. 

HaBiTAT. Canopy of trees in rain forest at lower elevations (150-350 m). 
DISTRIBUTION. Sri Lanka, Malaya, Philippines. 

ADDITIONAL SPECIMENS. Hale collections: 4 (47 058, 47 140), 6 (47 183), 9 (46 202). 


OBSERVATIONS. The apothecia are distinctly urceolate at maturity, although at earlier stages 
they are not strongly emergent. They also tend to occur in distinct clusters 1 cm or more 


across. 


12. Myriotrema fissurinum Hale sp. nov. 
(Fig. 12c) 


Thallus corticola, epiphloeodes, viridi-albicans, nitidus, aetate fissurinus, 7-12 cm latus; apothecia 
immersa, pro parte aggregata, 0'2-0'25 mm diametro, excipulo apice libero, columella nulla; 
ostiolum 0:05mm_ diametro; hymenium 70-150um altum; sporae fuscae, murales, 
10-12 x 27-32 um, 0-2 x 4-7 loculate, 4-8/ascus, I—. 


M. E. HALE 


Ae 


aa 


. (b) M. eminens (Hale 47 183). 
(c) M. fissurinum (Hale 46 227). (d) M. fluorescens (Hale 46 244). (e) M. frondosum (apothecia) 
(Hale 51 232). (f) M. frondosum (isidia) (Hale 51 232). (g) M. glaucescens (Hale 50 334). (h) M. 
glaucophaenum (Hale 47 148). (i) M. granulosum (Hale 47 092). See Fig. 7i for scale. 


THELOTREMATACEAE IN SRI LANKA 281 


Typus: Sri Lanka, Southern Province, Matara District, mossy forest above Enselwatta, elev. 
1100 m, 14 Feb. 1976, M. E. Hale 46 227 (US—holotypus; BM—isotypus). 


Thallus whitish to pale greenish gray or tan, 7-12 cm broad, shiny, smooth, fissured with 
age; cortex dense, 8-12 wm; algal layer continuous, 10-15 wm; medulla to 100 wm, with 
crystals; apothecia immersed in the medulla, often in masses, 0°2-0:25 mm diam, the exciple 
free at the tips, pale yellowish brown, partially filling the pore; pore nearly closed and 
darkening to open, 0°05mm diam; hymenium 70-150 um; spores brown, muriform, 
10-12 x 27-32 um,1-2 x 4-7 loculate, 4—-8/ascus, I—. 


CHEMISTRY. Stictic and constictic acids. 


HABITAT. Branches and lower trunk of trees in rain forest at mid and higher elevations 
(1100-2100 m). 


DISTRIBUTION. India, Sri Lanka. 


ADDITIONAL SPECIMENS. Hale collections: 5 (46 221), 1 (50 331). Thwaites collections: C. L. 
120, 122 (BM, H-Nyl. 3885, 3886, PDA, S, UPS) (as ‘Lgt. 82. Thelotrema glaucopallens’), C. 
L. 12 (BM, H-Nyl. 3874, PDA, S, UPS), C. L. (BM, PDA) (as ‘Lgt. 83. Thelotrema com- 
punctum Sm..’). India: Tamil Nadu, 3 km § Naduvattam, Hale 50 835, 50 844, 50 845, (US); 
Kerala, Devicolam, Hale 46 568 (US), Munnar, Hale 47 344, 47 355 (US). 


OBSERVATIONS. This species is closely related to M. subcompunctum (Nyl.) Hale, which I 
reported erroneously from Dominica (Hale, 1974a : 42). After more critical examination of 
type material in PC, I would now delimit M. subcompunctum as having a dull, weakly 
organized and porous cortex about 20 wm thick, a medulla to 150 wm thick arranged in 
vertical blocks with large crystalline inclusions, and spores 15-17 um long. I have also 
discovered two synonyms, Leptotrema diffractum Miill. Arg. from Australia (type: Bailey in 
G) and L. inclusum Zahlbr. from Japan (type: Faurie 5147 in W). Very close and perhaps 
representing additional synonyms are L. polycarpum Mill. Arg. from Australia (type: 
Knight in G) and L. polyporum Riddle from the West Indies (type: Jennings 229a in FH). I 
have not yet redetermined the Dominican collection. 

The material which I am describing here as M. fissurinum differs subtly but significantly in 
having a shiny, dense cortex 15 wm thick, a thick, continuous algal layer, and a medulla 
60-120 um thick with small grainy crystals, not organized in vertical arrays. The spores are 
also consistently larger than in M. subcompunctum, 21-32 um long. It is now known from 
southern India and Sri Lanka, generally above 1000 m elevation. 

Leighton had Nylander confirm the identification of ‘7. glaucopallens’. Notwithstanding, 
Thelotrema glaucopallens Nyl. (see Hale, 1978a:43 and Salisbury, 1971b:274) has 
colourless spores and an exfoliating cortex. While quite common in neighbouring India, I did 
not collect the species in Sri Lanka. Leighton also misidentified C. L. 12 and C. L. 129 as ‘T. 
compunctum Sm.’, which does occur in Sri Lanka but is not related at all to M. fissurinum. 


13. Myriotrema fluorescens Hale sp. nov. 
(Fig. 12d) 

Thallus corticola, epiphloeodes, nitidus, continuus, 6—20 cm latus; apothecia immersa, rotundata sed 
aetate elongata, 0°5-1‘(0 mm diametro, margine thallino suberecto, excipulo libero; ostiolum 
rotundatum vel elongato-fissum, 0°2-0°5 mm lato; hymenium 45-65 zm altum; sporae incolores, 
murales 4-8 x 12-15 wm, 1-2 x 4 loculatae, I+ coeruleae. 

Typus: Sri Lanka, Southern Province, Matara District, mossy forest above Enselwatta, elev. 
1100 m, 14 Feb. 1976, M. E. Hale 46 244 (US—holotypus; BM—isotypus). 


Thallus greenish mineral gray, 6-20 cm broad, shiny, continuous; cortex dense, 30-40 um, 
with aculeate hyphae, splitting into numerous sheets; algal layer continuous, 15 um; 
medulla 150-250 um with incorporated cortical sheets; apothecia immersed to semi- 
emergent, often in lirelliform-like clusters, the thalline rim suberect, apically crumbling with 
age and almost sorediate, exciple free apically, pale reddish brown, partially filling the pore; 


282 M. E. HALE 


pore gaping, round to slit, 0°2-0°5 mm wide; hymenium 45-65 yum; spores colourless, 
muriform, 4-8 x 12-15 wm, 1-2 x 4 loculate, I+. 


CHEMISTRY. Lichexanthone. 
HaBitAtT. Lower trunks of trees at low to mid elevations (300-1100 m). 
DISTRIBUTION. Sri Lanka. 


ADDITIONAL SPECIMENS. Hale collections: 1 (50 211, 50 218), 2 (51 124, 51 144, 51 170), 4 
(47 152). Thwaites collections: Lgt. 192 (BM, S) (as ‘Let. 192. Trypethelium schizostomum’). 


OBSERVATIONS. The elongate apothecia, thick, exfoliating cortex, and presence of 
lichexanthone mark this as an extraordinary species. It may be one of the possible bridges to 
the Graphidaceae, although the excipular structure is thelotremoid. 

When Thwaites divided the material of C. L. 121 (Trypethelium schizostomum), the 
Leighton set included only the species which I have determined as Ocellularia fissa (see 
below). Duplicates in BM and S are M. fluorescens. 


14. Myriotrema frondosum Hale sp. nov. 
(Fig. 12e, f) 


Thallus corticola, epiphloeodes, viridi-albus, 4-6 cm latu, crasse isidiatus, isidiis squamuloso-divisis; 
apothecia immersa, c. 0°2 mm diametro, excipulo libero, incolorato, columella nulla; ostiolum 
rotundatum, 0:05 mm diametro; hymenium | 15-125 wm altum; sproae incolores, murales, octonae, 
9-10 x 21-24 um, 1-2 x 5-6 loculatae, I+ coeruleae. 

Typus: Sri Lanka, Sabaragamuwa Province, Ratnapura District, Halwathura-Kanda, near 
Rassagala, elev. 900 m 15 Mar. 1978, M. E. Hale 51 232 (US—holotypus). 


Thallus light greenish gray, 4-6 cm broad, shiny, deeply fissured, coarsely isidiate, the isidia 
turning into flattened squamule-like excrescences 1 mm wide ‘and 1-1°5 mm high; cortex 
dense, 10-12 wm, with aculeate hyphae, splitting somewhat; algal layer 15 wm; medulla 
30-50 um; apothecia sparse and inconspicuous, immersed in the medulla, 0°2 mm 
diam, the exciple colourless, free apically, in part filling the pore; pore round, 0°05 mm 
diam; hymenium 115-125 um; spores colourless, muriform, 9-10 x 21-24 wm, 1-2 x 5-6 
loculate, I+. 


CHEMISTRY. Psoromic acid. 
HABITAT. Tree trunks at mid elevations in mossy forest (900 m). 
DISTRIBUTION. Sri Lanka. 


OBSERVATIONS. This species produces clumps of unique, isidioid, subfoliar structures. I first 
thought that I had collected a Myriotrema species overgrowing a liverwort, but the growths 
are entirely lichenized. It isa member of the ‘Thelotrema album’ group. 


15. Myriotrema glaucescens (Nyl.) Hale in Mycotaxon 11 : 133 (1980). 
(Fig. 12 g) 


Thelotrema glaucescens Nyl. in Annls Sci. nat. (Bot.) IV, 19 : 332 (1863). Type: USA, Louisiana, Dr 
Hale s.n. (FH-Tuck.—lectotype). 

Leptotrema glaucescens (Nyl.) Mill. Arg. in Flora, Jena 65 : 499(1882). 

IcONES. Redinger, 1936 : 108 fig. 68. 


Thallus whitish mineral gray, 3-10 cm broad, dull, fissuring and sometimes bulging out; 
cortex lacking; medulla compact, cellular, 150-200 wm with algae between vertical blocks; 
apothecia immersed in the medulla, round to angular, 0°2-0°3 mm long, often massed, the 
exciple fused, the disc pulling away from the wall; disc blackish to white pruinose; 
hymenium 90-130 4m; spores brown, muriform, 7-10 x 13-16 um, 1-2 x4 loculate, 
uniseriate, I—. 


THELOTREMATACEAE IN SRI LANKA 283 
CHEMISTRY. Stictic and constictic acids. 
Hasitatv. Planted trees at mid or higher elevations (SO0-2100 m). 
DISTRIBUTION. United States, West Indies, Central and South America, India, Sri Lanka. 


ADDITIONAL SPECIMENS. Hale collections: 11 (50 334), 13 (50 049). Thwaites collections: C. 
L. 119 (BM, H-Nyl. 3918, S, UPS) (as ‘Lgt. 74. Th. glaucescens’). Alutnuwara, Alston 1952 
(PDA). 


OBSERVATIONS. This species can be recognized easily by the crowded, black, angular 
apothecia and the presence of stictic acid. The thallus surface is dull and grainy since 
medullary hyphae are exposed directly (Fig. Sa). Leighton correctly identified it by 
comparison with Wright’s Lich. Cubae 151 and 152. 


16. Myriotrema glaucophaenum (Krempelh.) Hale in Mycotaxon 11 : 133 (1980). 
(Fig. 12h) 


Thelotrema glaucophaenum Krempelh. in Nuovo G. bot. ital. 7:19 (1875). Type: Sarawak, 1866, 
Beccari 92 (M—lectotype; G—isolectotype). 

Ocellularia glaucophaena (Krempelh.) Zahlbr., Cat. Lich. Univers. 2 : 591 (1923). 

Icones. Hale, 1978a : 5, 22 fig. Ir (lectotype) and fig. 6b. 


Thallus pale greenish to yellowish mineral gray, 8-10 cm broad, shiny, continuous; cortex 
dense, 10-15 wm, with aculeate hyphae, splitting internally with some exfoliation; algal 
layer continuous, 10-12 wm; medulla 10-90 um with crystals; apothecia barely emergent, 
0:3-0:4 mm diam, the exciple fused, colourless, columella lacking; pore round, 0°1-0:15 mm 
diam; hymenium c. 60 um; spores colourless, transversely septate, 3-S x 6-12 um, 34 
loculate, I+. 


CHEMISTRY. Psoromic acid. 
HABITAT. Lower trunk and lianas in rain forest at low elevations (250-350 m). 


DISTRIBUTION. United States, West Indies, Central America, India, Sri Lanka, Java, 
Sarawak, Solomon Islands. 


ADDITIONAL SPECIMENS. Hale collections: 4 (47 059, 47 061, 47 062, 47 148), 7b (51 247). 


OBSERVATIONS. The material from Sri lanka is very close to the type specimen in all respects. 
It can be usually be distinguished from M. terebratulum by the raised apothecia and thinner 
thallus. This complex, however, is a difficult one and the species limits are unsettled (see 
Hale, 1978a: 21). 


17. Myriotrema granulosum (Leighton) Hale in Mycotaxon 11 : 133 (1980). 
(Fig. 121) 

Ascidium granulosum Leighton in Trans. Linn. Soc. Lond.27 : 171 (1870). [Non Ocellularia granulosa 
(Tuck.) Zahlbr. ex anno 1923]. Type: Sri Lanka, south of the island, Thwaites C. L. 127 
(BM—lectotype; G, PC—isotectotypes). 

Ocellularia leightonii Zahlbr., Cat. Lich Univers. 2 : 593 (1923). Based on Ascidium granulosum. 


Thallus greenish mineral gray, 7-12 cm broad, coarsely verrucose and deeply fissured; 
cortex dense, 10-12 wm; algal layer continuous, 12 wm; medulla 15 um, pale yellowish, 
mostly hypophloeodal; apothecia numerous, emergent, 0°3-0°5 mm diam, slightly con- 
stricted basally and appearing urceolate at maturity, the exciple fused, reddish brown, free 
apically, columella lacking; pore round, 0°05 mm diam, distinctly white annulate; 
hymenium 280-300 um; spores colourless, transversely septate, 15-18 x 90-150 wm, 14-16 
loculate, rarely with a longitudinal septum in the end loculae, 1—2/ascus, I+. 


CHEMISTRY. Unknown P+ substance (‘leightonii’ unknown). 


HABITAT. Saplings, mid bole and canopy in rain forest at lower elevations (200-300 m). 


284 M. E. HALE 


DISTRIBUTION. Sri Lanka. 


ADDITIONAL SPECIMENS. Hale collections: 4 (47 060, 47 092, 50 487, 50 512), 6 (47 173, 47 
187), 7a (50 152). Thwaites collections: C. L. 127 (BM, G, H-Nyl. 3870, PC) (as ‘Lgt. 100, 
Ascidium granulosum’, C. L. 260 (PDA). 


OBSERVATIONS. This endemic species is remarkable for the almost urceolate apothecia, 
coarsely verruculose thallus, large spores, and the unique chemistry and pale yellow 
medullary pigment. The P+ unknown substances is probably an undescribed depsidone. 


18. Myriotrema hartii (Mull. Arg.) Hale in Mycotaxon 11 : 133 (1980). 
(Fig. 13a, b) 


Thelotrema hartii Mill Arg. in Flora, Jena 69 : 311 (1886). Type: Jamaica, Gordontown, Hart s.n. 
(G—lectotype; BM, NY (as no. 36)—isolectotypes). 


Thallus pale greenish white, c. 8 cm diam, shiny, deeply fissured, coarsely isidiate, the isidia 
0-2 mm diam and to 0°8 mm high; cortex dense and irregularly thickened, 8-15 wm, with 
aculeate hyphae; algal layer continuous, 15-20 um; medulla 10-60 wm with crystals; 
apothecia numerous, immersed in the medulla, 0:2-0°3 mm diam, the thalline rim becoming 
raised with age, exciple apically free, pale yellowish, visible through the pore, columella 
lacking; pore round, 0°1-0°15 mm diam, white rimmed; hymenium c. 40 um; spores 
colourless, muriform, 6 x 13-15 wm, 1-2 x 4 loculate, I+. 


CHEMISTRY. Psoromic acid. 

HaBITAT. Branches in open forest at high elevations (2100 m). 
DISTRIBUTION. West Indies, Central and South America, Sri Lanka. 
ADDITIONAL SPECIMENS. Hale collections: 10 (50 270). 


OBSERVATIONS. When describing Thelotrema hartii, Miller Argoviensis did not mention any 
‘isidia’, although these structures are quite distinct and present in the type material. It is 
indeed very close to M. concretum (Fée) Hale, as he suggested, and may prove eventually to 
be merely a morphological variant. One often collects coarsely isidiate but sterile thalli in 
the tropics closely resembling M. hartii. Since M. hartii often has very sparsely developed 
apothecia, it may well be widely distributed as these sterile crusts. 


19. Myriotrema masonhalei (Patw. & C. Kulk.) Hale in Mycotaxon 11 : 134 (1980). 
(Fig. 13c) 
Thelotrema masonhalei Patw. & C. Kulk. in Norw. J. Bot. 24: 128 (1977). Type: India, Maharashtra, 


Amboli, A. V. Prabhu & M. B. Nagarkar 74.2262 (AMH—holotype). 

ICONES. Patwardhan & Kulkarni, 1977 : 129 fig. 4. 

Thallus whitish ashy gray, 6-12 cm broad, shiny, smooth to verruculose, fissured with age; 
cortex dense, irregularly thickened, 8-15 wm, with some aculeate hyphae; algal layer 
continuous, 10 um; medulla 0-60 wm with crystals, mostly hypophloeodal; apothecia 
inconspicuous, immersed deeply in the periderm, 0°5-O°7 mm diam, exciple fused, 
colourless, columella lacking; pore round, 0:05-0:08 mm diam, white rimmed; hymenium 
180-350 um; spores colourless, muriform, 30-45 x 150-250 wm, with numerous locules, 
1/ascus, I+. 


CHEMISTRY. Psoromic acid. 


HABITAT. Lianas, saplings, and exposed roots, rarely into the canopy, in rain forest over 
broad elevations (150-2100 m). 


DISTRIBUTION. India, Sri Lanka. 


THELOTREMATACEAE IN SRI LANKA 285 


Fig. 13 Species of Myriotrema. (a) M. hartii (apothecia) (Hale 50 270). (b) M. hartii (isidia) (Hale 
50 270). (c) M. masonhalei (Hale 46 341). (d) M. mastarion (Hale 51 039). (e) M. microporum 
(Hale 46 217). (f) M. microstomum (Hale 46 223). (g) M. minutulum (Hale 46 358). (h) M. 
minutum (Hale 47 033). (1) M. multicavum (Hale 46 150). See Fig 71 for scale. 


286 M. E. HALE 


ADDITIONAL SPECIMENS. Hale collections: | (50 214, 50 219, 50 252 50 260), 2 (51 164), 3a 
(46 290, 46 341, 46 347, 46 348), 4 (47 071, 47 111, 47 128, 50 382), 6 (47 196), 7a (50 178), 
7b (51 024, 51 082, 51 250), 8 (47 027), 9 (46 215), 10 (50 264, 50 288). 


OBSERVATIONS. This is the third most common Myriotrema in Sri Lanka. It is also common 
in the evergreen monsoon forests of the Western Ghats in India. The deeply immersed 
apothecia are very inconspicuous, their presence indicated only by the tiny, white rimmed 
pore. However, once the large spores are found and the chemistry determined, the species is 
easily identified. 


20. Myriotrema mastarion Hale sp. nov. 
(Fig. 13d) 


Thallus corticola, epiphloeodes, nitidus, continuus, 8-10 cm latus; apothecia profunde immersa, 
0:3-0°-4 mm diametro, excipulo connato, columella nulla; ostiolum rotundatum, 0°08—-0°1 mm 
diametro; hymenium 120-180 wm altum; sporae fuscae, octonae, transversim septatae, 5-9 x 15- 
21 um, 4 loculatae, I—. 

Typus: Sri Lanka, Western Province, Kalatura District, Morapitiya, Ambalam-pola, elev. 300 m, 
13 Mar. 1978, M. E. Hale 51 039 (US—holotypus; BM—isotypus). 


Thallus ashy greenish or tannish gray, 8-10 cm broad, shiny, continuous, faintly grainy; 
cortex loosely organized and irregularly pored, 12 um; algal layer 10-20 um, more or less 
interrupted by crystals; medulla 15-45 wm, sometimes faintly yellowish, with crystals; 
apothecia inconspicuous, deeply immersed in periderm layers, 0°3-0°-4 mm diam, the 
exciple fused, reddish brown, columella lacking; pore round, 0:08-0°1 mm diam, the rim 
raised, tannish white; hymenium 120-180 um; spores brown, inconspicuous, deeply 
immersed in periderm layers, 0°3-0'4 mm diam, the exciple fused, reddish brown, columella 
lacking; pore round, 0:08-0:1 mm diam, the rim raised, tannish white; hymenium 120-180 
sum; spores brown, transversely septate, S—9 x 15-21 wm, 4 loculate, I—. 


CHEMISTRY. Psoromic and norpsoromic acids. 


HasitaT. Trunks along open trails and in canopy in rain forest at low to mid elevations 
(150-850 m). 


DISTRIBUTION. Sri Lanka. 
ADDITIONAL SPECIMENS. Hale collections: 2 (51 233), 4 (47 159, 50 520), 9 (46 207). 


OBSERVATIONS. This species closely resembles M. andamanicum (see above) in apothecial 
structure and chemistry; only the spores are different (muriform in M. andamanicum). 


21. Myriotrema microporum (Mont.) Hale in Mycotaxon 11 : 134 (1980). 
(Fig. 13e) 


Thelotrema microporum Mont. in Annis Sci. nat. (Bot.) Ill, 10 : 130 (1848). Type: Java, Junghuhn, 
Lichen. Javan. 143 (PC—lectotype; FH-Tuck., G, H-Nyl., L—isolectotypes). 

Thelotrema crassulum Nyl. in Annls Sci. nat. (Bot.) IV, 11 : 258 (1859). Type: Bourbon [= Réunion], 
Biovin s.n. (PC—lectotype; G, H-Nyl. 22652—isolectotypes). 

Ocellularia micropora (Mont.) Mill. Arg. in Flora, Jena74 : 112 (1891). 

Ocellularia crassula (Nyl.) Zahlbr., Cat. Lich. Univers. 2 : 588 (1923). 


Thallus light greenish to whitish mineral gray, 8-15 cm broad, thick and deeply fissured, 
smooth, shiny; cortex barely formed, about 5 wm, with aculeate hyphae; medulla densely 
cellular, columnar, 200-300 um thick, the algae between vertical arrays in layers 10-25 wm; 
apothecia numerous, immersed in the medulla, 0:2-0°3 mm diam, the exciple more or less 
free apically, pale reddish, visible through the pore, columella lacking; pore round, 
0:07-0'1 mm diam with a slightly raised whitish rim; hymenium 55-60 wm; spores 
colourless, transversely septate, 4 x 8-11 wm, 3-4 loculate, I+. 


THELOTREMATACEAE IN SRI LANKA 287 


CHEMISTRY. Psoromic and norpsoromic acid. 
HaBitAT. Canopy branches in rain forest at mid higher elevations (300-2000 m). 


DISTRIBUTION. Mascarene Islands, India, Sri Lanka, Malaya, Philippines, Java, Sabah, 
Taiwan, Japan, Solomon Islands, Australia. 


ADDITIONAL SPECIMENS. Hale collections: 4 (47 083), 5 (46 271), 10 (SO 287). Thwaites 
cocetions: C. L. 142 (PDA). C. L. 143 (BM, H-Nyl., 3852, S, UPS) (as ‘Let. 75, T. 
leptoporum’) (material in BM and S admixed with Myriotrema olivaceum). 


OBSERVATIONS. Salisbury (1978 :411) followed Redinger (1936: 48) in considering this 
species as a synonym of Thelotrema (Myriotrema) olivaceum. As far as I can determine, 
these two species are not closely related, differing in both chemistry and morphology. 
Mpyriotrema olivaceum has a discrete, thick cortex (15-18 wm) and on average a thinner 
medulla (180-200 wm) than M. microporum. Nylander suggested that his Thelotrema 
crassulum might be combined 7. microporum (he had undoubtedly seen the Montagne type 
in PC) and later (1869:70) questioned whether 7. microporum was distinct from 
Myriotrema album. As | have shown above, M. album lacks any chemistry and has a much 
thinner thallus. The deep fissuring characteristic of M. microporum was first noted by 
Harmand (1912 : 36) as one way to separate it from M. album. 

As far as the Sri Lankan flora is concerned. M. microporum might be confused with M. 
terebratulum, which has identical chemistry but a much thicker cortex (about 30 wm) and 
little fissuring. I have not yet identified M. microporum from the New World. 

Leighton’s determinations of ‘Thelotrema microporum’, none of them checked by 
Nylander, are all incorrect. Some of the material is M. thwaitesii (see below), some M. 
rugiferum (below). At the same time he misidentified M. microporum as ‘Thelotrema 
leptoporum Nyl.=Wright’s Lich. Cubae, 128!. Thelotrema leptoporum Nyl. (see Hale, 
1978a: 23) is a New World species with hypoprotocetraric acid and lichexanthone, a 
synonym of Myriotrema glauculum (Nyl.) Hale. 


22. Myriotrema microstomum (Mill. Arg.) Hale in Mycotaxon 11 : 134 (1980). 
(Fig. 13f) 


Thelotrema microstomum Mill. Arg. in Flora, Jena 74: 113 (1891). Type: Japan, Mt. Tosa, Miyoshi 
(G—lectotype; FH, TI—isolectotypes). 

Thelotrema microstomum var. formosanum Zahlbr. in Reprium Spec. nov. Regni veg. 33: 24 (1933). 
Type: Taiwan, Regechi, Asahina 233 (W—lectotype). 


Thallus pale greenish ashy, 8—12 cm broad, shiny, irregularly cracked with age; cortex dense, 
10-15 wm, with aculeate hyphae, splitting internally; algal layer continuous, 10-15 wm; 
medulla 5-20 wm with large crystals; apothecia rather numerous, emergent, solitary or 
clustered in 2’s and 3’s, 1-1-8 mm diam, the exciple fused, colourless; columella lacking; 
pore round, 0°1-0°18 mm diam, somewhat depressed, white rimmed and with white 
markings radiating outward over the amphithecium; hymenium 160-320 um; spores colour- 
less, muriform, 30-45 x 120-150 wm, with numerous locules, l/ascus, I+. 


CHEMISTRY. Fumarprotocetraric acid with traces of protocetraric acid. 
HABITAT. Lower trunks of trees in mid elevation mossy forest (850-1100 m). 
DISTRIBUTION. Japan, Taiwan, Sri Lanka. 


ADDITIONAL SPECIMENS. Hale collections: 2 (51 098, 51 126, 51 137, 51 214), 5 (46 223, 46 
228, 46 231, 46 236, 46 258, 46 261). Thwaites collections: C. L. 97 (BM), C. L. 257 (PDA) 
(mixed with Ocellularia pertusariiformis), Lgt. 96 (BM) (as ‘T. pertusariiforme’). 


OBSERVATIONS. The Sri Lankan material is very close to the Japanese type (it is identical to 
var. formosanum) in pore configuration, although the rim around the pore is not distinctly 
raised. The species is unusual because of the large, noncarbonized apothecia, large spores, 


288 M. E. HALE 


and chemistry. Related M. cinereum (Mill. Arg.) Hale from Japan has fumarprotocetraric 
acid, but the apothecia are smaller and less emergent. Another species producing 
fumarprotocetraric acid, M. secernendum (Harm.) Hale, from New Caledonia has immersed 
apothecia and small spores (26 wm long). 


23. Myriotrema minutulum (Hale) Hale in Mycotaxon 11 : 134 (1980). 
(Fig. 13g) 
Ocellularia minutula Hale in Smithson. Contr. Bot. 38:24 (1978). Type: Panama, Canal Zone, 11 


Feb. 1974, M. E. Hale 43 341 (US—holotype). 
Icons. Hale, 1978a : 22 fig. 6h. 


Thallus pale greenish ashy, 8-12 cm broad, shiny, continuous, smooth to rugulose; cortex 
dense, 10-12 wm, with sparse aculeate hyphae; algal layer continuous, 6-10 wm; medulla 
8-12 um with numerous crystals, mostly hypophloeodal; apothecia common but not 
conspicuous, immersed deeply in periderm tissue, 0°2-0°3 mm diam, the exciple fused, 
colourless; pore round, 0:05-0:07 mm diam, slightly depressed and white rimmed; 
hymenium 80-100 mum; spores colourless, transversely septate, 5-6 x 15-24 wm, 5-6 
loculate, I+. 


CHEMISTRY. Psoromic acid. 

HABITAT. Lower trunk to mid bole in rain forest at low elevations (300 m). 
DISTRIBUTION. Central America, Sri Lanka. 

ADDITIONAL SPECIMENS. Hale collections: 3a (46 358, 46 369), 7b (51 086, 51 093). 


OBSERVATIONS. The apothecia are deeply immersed and inconspicuous. The species is 
probably related to M. masonhalei, which has larger apothecia and muriform spores. 


24. Myriotrema minutum (Hale) Hale in Mycotaxon 11 : 134 (1980). 
(Fig. 13h) 


Ocellularia minuta Hale in Mycotaxon7 : 379 (1978). Type: Sri Lanka, Southern Province, Morawaka 
Hill, 15 Feb. 1976, M. E. Hale 47 033 (US—holotype; BM, PDA—isotypes). 
ICONES. Hale, 19786 : 384 fig. 4. 


Thallus greenish gray, 8-l10cm broad, thick, shiny, broadly fissured; cortex dense, 
10-15 wm, with aculeate hyphae, splitting internally; algal layer continuous, 10-15 um; 
medulla 150 um or more with incorporated cortical sheets and dense crystal masses; 
apothecia very numerous, immersed in the medulla, 0:1-0°2 mm diam, the exciple fused, 
colourless; pore round, 0:05 mm diam, white rimmed; hymenium 60-70 um; spores colour- 
less, transversely septate, 4-6 x 18-20 wm, 4-5 loculate, I+. 


CHEMISTRY. Fumarprotocetraric acid. 
HABITAT. Lower trunks in rain forest at lower elevations (300-800 m). 
DISTRIBUTION. Sri Lanka, Sarawak. 


ADDITIONAL SPECIMENS. Hale collections: 2 (51116), 7a (50 315). Sarawak: Mt Matang, Hale 
30 768 (US). 


OBSERVATIONS. This is one of the few species in the genus with fumarprotocetraric acid. 
Other diagnostic features are the very small immersed apothecia and the thick, internally 
splitting cortex which is successively incorporated in the medulla as new tissue forms. 


25. Myriotrema multicavum Hale sp. nov. 
(Fig. 131) 


Thallus corticola, epiphloeodes, crassus, fere tumulatus, fissurinus, friabilis, cinereo-albus vel pallide 
viridis, 8-12 cm latus; apothecia numerosa, solitaria, profunde immersa, 0°2-0°3 mm diametro, 


THELOTREMATACEAE IN SRI LANKA 289 


excipulo apice libero, columella nulla; ostiolum rotundatum, c. 0°! mm diametro; hymenium 
100-120 wm altum; sporae incolorae, transversim septatae, octonae, male evolutae, 4 x 8 wm, 3-4 
loculate, I+ coeruleae. 

Typus: Sri Lanka, Southern Province, Galle District, Kanneliya Forest Reserve, elev. 150 m, 16 
Feb. 1976, M. E. Hale 46 150 (US—holotypus; BM—isotypus). 


Thallus pale greenish ashy, 8—12 cm broad, shiny, grainy, thick and deeply fissured, friable; 
cortex very thin, c. 5 zm and sometimes not distinguishable from the medulla, with long 
aculeate hyphae; medulla densely cellular, 200-300 um thick in vertically oriented arrays, 
with dense crystals, the algae lying between columns; apothecia numerous, immersed in the 
medulla, 0°2-0°3 mm diam, the exciple free at the tips, colourless, visible through the pore; 
pore round, 0°! mm diam; hymenium 100-120 um; spores colourless, transversely septate, 
4x 8 wm, 3-5 loculate, poorly developed, I+. 


CHEMISTRY. Psoromic acid. 

HasirTAT. Canopy of trees in rain forest at lower elevations (150-300 m). 
DISTRIBUTION. Sri Lanka. 

ADDITIONAL SPECIMENS. Hale collections: 4 (47 147). 


OBSERVATIONS. This new species appears superficially to be close to M. microporum or M. 
terebratulum, but it can be separated by the grainy surface and the columnar, crystal-filled 
medulla (Fig. Sc). 


26. Myriotrema nuwarense Hale sp. nov. 
(Fig. 14a) 


Thallus corticola, epiphloeodes, albo-cinereus, rimosus, usque ad 12 cm latus; apothecia immersa vel 
pauce emergentia, 0°2-O-mm diametro, excipulo libero, crasso, incolorato, columella nulla; 
ostiolum rotundatum, 0°! mm diam; hymenium 200-210 um altum; sporae fuscae, murales, 
octonae, 12-14 x 45-60 wm, 1-2 x 7-9 loculatae, 2 : nae, I—. 

Typus: Sri Lanka, Central Province, Nuwara Eliya District, Forest Reserve near Nuwara Eliya, 
elev. 2100 m, 21 Mar. 1978, M. E. Hale 50 272 (US—holotypus; BM—isotypus). 


Thallus whitish to tannish gray, to 12 cm broad, dull, fissured; cortex dense, 10-15 wm; algal 
layer continuous, 15-18 wm; medulla 60-120 wm, with some crystals; apothecia numerous, 
immersed in the medulla, slightly raised at maturity, 0°2-0°3 mm diam, exciple free apically, 
thickened, colourless, partially filling the pore; pore round, c. 0°! mm diam; hymenium 
200-210 um; spores brown, muriform, 12-14 x 45-60 wm, 1-3 x 7-9 loculate, 2-4/ascus, 
I-. 


CHEMISTRY. Stictic and constictic acids. 
HABITAT. Small branches of trees at high elevations (2100 m). 
DISTRIBUTION. Sri Lanka. 


OBSERVATIONS. It is with great hesitation that I describe another stictic acid-containing 
species of Myriotrema. However, M. nuwarense differs fairly strongly from others in the 
group in having immersed but conspicuous apothecia with an apically free exciple which 
fills the pore in part. The closest relative is probably M. fissurinum (see above). 


27. Myriotrema olivaceum Feée, Essai Crypt : 103 (1824). 
(Fig. 14b) 


Type: ?South America, on Bonplandia trifoliata (G—lectotype; M, PC-Mont.—isolectotypes). 
Ocellularia olivacea (Fée) Miill. Arg. in Mém. Soc. Phys. Hist. nat. Genéve 29(8) : 7 (1887). 
Icongs. Redinger, 1936 : 48 fig. 24.—Hale, 1974a : 7, 23 figs Sc and 12c (lectotype).—Hale, 1978a : 22 
fig. 61.—Salisbury, 1978 : 410 fig. 4. 


Thallus pale greenish ashy gray, 10-15 cm diam, smooth, shiny, continuous to fissured with 


290 M. E. HALE 


Oey « 
ao 


Fig. 14 Species of Myriotrema. (a) M. nuwarense (Hale 50 272). (b) M. olivaceum (Hale 50 145). (c) 
M. polytretum (Hale 50 238). (d) M. porinaceum (Hale 51 142). (e) M. protoalbum (Hale 51 044). 
(f) M. rugiferum (Hale 47 058). (g) M. santessonii (Wheeler 12 471 in US). (h) M. subconforme 
(Hale 50 510). (i) M. terebrans (Hale 47 202). See Fig. 71 for scale. 


THELOTREMATACEAE IN SRI LANKA 291 


age; cortex dense, 15-18 wm, with conspicuous aculeate hyphae; algal layer continuous, 
15 um; medulla 180-200 um with penetrations to 400 wm in eroding pockets of periderm; 
apothecia very numerous, immersed in the medulla, 0:2-0°3 mm diam, exciple fused, often 
free apically, colourless, pore round, 0:05-0:08 mm diam, becoming white rimmed and 
somewhat raised; hymenium 60-80 um; spores colourless, transversely septate, 4-6 x 
10-15 um, 4 loculate, I+. 


CHEMISTRY. “Olivacea’ unknowns (2 P— spots). 
HABITAT. Canopy branches in rain forest at low to mid elevations (150-800 m). 


DISTRIBUTION. West Indies, Central and South America, Sri Lanka, New Caledonia, 
Marquesas Islands. 


ADDITIONAL SPECIMENS. Hale collections: 1 (50 226), 3a (46 291, 46 294, 46 352, 46 354), 4 
(50 375), 6 (47 056, 47 190, 47 197, 47 203, 47 213), 7a (50 145), 7b (51 003, 51 029, 51 059, 
51 073, 51 084, 51 253), 8 (47 007), 9 (46 138, 46 141, 46 145, 46 184, 46 185, 46 187, 47 
007). Thwaites collections: Lgt. 75 as ‘Thelotrema leptoporum’ (BM, S). Matara District: E of 
Gandara, Santesson 25 735 (S). 


OBSERVATIONS. This canopy-dwelling species is essentially pantropical. The chemistry is 
distinctive although the identity of the two presumed depsidone components is still 
unknown. It is most often compared with M. album (e.g. Miller Argoviensis, 1887 : 7), 
which lacks any chemistry and has a much thinner thallus. The alleged olivaceous colour is 
probably taken from the type; fresh material is whitish or pale greenish mineral gray. 
Furthermore, as mentioned above, M. microporum cannot be considered a synonym of this 
species since it contains psoromic acid (P+ yellow). Another canopy species, M. tere- 
bratulum, should be recognized as distinct on the same basis. While M. olivaceum is wide- 
spread in the rain forest of Sri Lanka, it is curious that I found it only once in the extensive 
Weddagala logging area, showing how non-randomly the species may occur in a continuous 
forested area. 


28. Myriotrema polytretum Hale sp. nov. 
(Fig. 14c) 
Thallus corticola, epiphloeodes, viridi-cinereus, 6-8 cm latus; apothecia numerosa, emersa, 0-3 mm 
diam, excipulo apice libero; ostiolum rotundatum, 0-1 mm diametro; hymenium 40-45 um altum; 


sporae incolores, transversim septatae, octonae, 4 x 12 wm, 4 loculatae, I+ coeruleae. 
Typus: Sri Lanka , Sabaragamuwa Province, Kegalla District, elev. 800 m Maliboda to Theberton, 


16 Mar. 1978, M. E. Hale 50 238 (US—holotypus; BM—isotypus). 
Thallus light greenish mineral gray, 6-8 cm broad, shiny, continuous; cortex very dense, 
10 wm, with aculeate hyphae and some internal splitting; algal layer 15 um; medulla 
40-80 wm, with numerous crystals; apothecia common, immersed in the medulla, 
0:2-0:3 mm diam, the exciple fused but partially free at the tips, pale orange red; pore round, 
c. 01 mm diam; hymenium 40-45 um; spores colourless, transversely septate, 4 x 12 um, 4 
loculate, I+. 


CHEMISTRY. ‘Tl hwaitesii’ unknown (P+ red). 
Hapsitat. Tree trunks along trail in mossy forest (900 m). 
DISTRIBUTION. Sri Lanka, Malaya, Sarawak. 


ADDITIONAL SPECIMENS. Hale collections: Malaya, Selangor, 30 271 (US); Sarawak, Sibu, 
30 434 (US) 


OBSERVATIONS. Externally M. polytretum is very similar to M. album, but lam separating it 
largely on the basis of the unusual chemistry (M. album lacks chemistry) and the thicker 
thallus. 


292 M. E. HALE 


29. Myriotrema porinaceum (Mill. Arg.) Hale in Mycotaxon 11 : 134 (1980). 
(Fig. 14d) 


Thelotrema porinaceum Mill. Arg. in Nuovo G. bot. ital. 23: 130 (1891). Type: Japan, Awa, 
Miyoshi 17 (G—lectotype; FH—probable isolectotype). 


Thallus greenish mineral gray, 3-6 cm broad, dull and thin, continuous; cortex loosely 
organized, 10 wm; algal layer discontinuous, 10 wm; medulla to 10 ~m but mostly hypo- 
phloeodal; apothecia dispersed, semi-emergent but inconspicuous, 0°7-0°9 mm diam, the 
thalline rim lacking periderm cells, exciple fused, colourless; pore round, 0°05 mm diam, 
more or less white rimmed; hymenium 280-300 um; spores colourless, muriform, 
20-30 x 60-180 um, with numerous locules, 1/ascus, I+. 


CHEMISTRY. Norstictic acid. 

HABITAT. Lower trunk of trees along trails at mid elevations or lower (350-850 m). 
DISTRIBUTION. Sri Lanka, Japan, Australia. 

ADDITIONAL SPECIMENS. Hale collections: 2 (51 142), 4 (50 305). 


OBSERVATIONS. Norstictic acid as the main constituent is very rare in Myriotrema, known so 
far only in M. norsticticum (Hale) Hale and M. occultum (Eschw.) Hale from the New World; 
pantropical M. compunctum has stictic acid intermixed. Myriotrema porinaceum 1s a rather 
inconspicuous lichen since the sparsely developed, semi-emergent apothecia have a tiny 
pore and blend into the background of the thallus. Miiller Argoviensis had compared the 
species with ‘Thelotrema reclusum Nyl.’ (= Myriotrema reclusum (Krempelh.) Hale), a 
stictic acid-containing species from the Andaman Islands (see Hale, 1978a : 54). 


30. Myriotrema protoalbum Hale sp. nov. 
(Fig. 14e) 


Thallus corticola, epiphloeodes, continuous, obscure viridis, 4-8 cm latus; apothecia immersa, incon- 
spicua, 0°1-O0°2 mm diametro, excipulo connato; hymenium 30-40 wm altum; sporae incolores, 
transversim septatae, 3-7 x 8-18 um, 4 loculatae, I—. 

Typus: Sri Lanka, Western Province, Kalatura District, Ambalam-pole, Morapitiya, elev. 300 m, 
13 Mar. 1978, M. E. Hale 51 044 (US—holotypus; BM—isotypus). 


Thallus dark greenish gray, 4-8 cm broad, shiny, continuous; cortex very dense, | 1-14 wm, 
with short aculeate hyphae; algal layer continuous, 8-10 um; medulla 10-30 um, with 
crystals, mostly hypophloeodal; apothecia numerous but inconspicuous, 0°1-0:2 mm diam, 
immersed in part in the periderm, becoming slightly raised and darkening, the exciple fused, 
colourless; pore round, 0:01 mm diam, the rim whitish yellow; hymenium 30-40 um; spores 
colourless, transversely septate, 3-7 x 8-18 um, 4 loculate, I—. 


CHEMISTRY. No substances present. 

HABITAT. Lower trunk in rain forest at lower elevations (300-800 m). 
DISTRIBUTION. Sri Lanka. 

ADDITIONAL SPECIMENS. Hale collections: 1 (50 247), 2 (51 219). 


OBSERVATIONS. The tiny apothecia are sometimes arranged in rows. The thallus is very thin 
and dark compared with M. album to which it seems to be related. It also falls close to M. 
cinereoglaucescens, a larger species. 


31. Myriotrema rugiferum (Harm.) Hale in Mycotaxon 11 : 135 (1980). 
(Fig. 14f) 


Thelotrema rugiferum Harm. in Bull. Séanc. Soc. Sci. Nancy Ill, 13 : 44 (1912). Type: New Caledonia, 
Pionniero s.n. (DUKE—lectotype; S—isolectotype). 


THELOTREMATACEAE IN SRI LANKA 293 


Thallus ashy greenish white, 7-15 cm broad, shiny, continuous; cortex dense, 15 wm, with 
well developed aculeate hyphae, splitting and exfoliating; algal layer continuous, 10-15 wm; 
medulla rather thick, 50-150 zm, with numerous crystals; apothecia numerous, immersed 
in the medulla, 0°3-0-4 mm diam, the exciple free apically, colourless, often visible through 
the pore; pore round, 0°1-0°:15 mm diam, white rimmed; hymenium 60-90 wm; spores 
colourless, muriform, 4-6 x 9-20 wm, 1-3 x 3-5 loculate, I+. 


CHEMISTRY. Psoromic acid with or without norpsoromic acid. 


HABITAT. Mainly the lower trunks but also into the canopy in rain forest at all elevations 
(150-2100 m). 


DISTRIBUTION. India, Sri Lanka, Philippines, New Caledonia, Solomon Islands. 


ADDITIONAL SPECIMENS. Hale collections: 1 (50 232, 50 246), 2 (51 222), 3a (46 364), 4 
(47 085, 47 150, 47 160, 50 373, 50 377, 50 428, 50 459, 50 504), 5 (46 232, 46 243, 
46 256, 46 257, 46 262), 7b (50 310), 10 (50 279, 50 290). Thwaites collections: CL. 11 (BM, 
H-Nyl. 3889, S, UPS), C. L. 142 (BM, PDA) (as ‘Lgt. 79. 7. concretum’), C. L. 144 (BM, 
mixed with Myriotrema thwaitesii) as ‘Lgt. 76. T. microporum’). 


OBSERVATIONS. In view of Salisbury’s (1978 :412) recent treatment of Thelotrema (Myrio- 
trema) clandestinum, | re-examined all of the material in US, some 52 collections. I 
confirmed that Fée’s type in G has muriform spores (see Hale, 1978a : 41). I also found that 
there are two different populations. One of these, which I presume to be the typical 
population, occurs only in the New World and has a dull surface without exfoliating layers, 
few crystals in the medulla, and a generally ‘hidden’ exciple, that is the tips of the exciple 
do not protrude into the pore as an inner ring. The other population, which occurs through- 
out the Old World and in Panama, has an exfoliating cortex, numerous large crystals in the 
medulla, and a conspicuous exciple visible in the pore. This population is M. rugiferum. 

Leighton used ‘Lindig, Lich. Nov. Gran. 59’ to identify C. L. 11 and C. L. 142 as 
Thelotrema concretum Fée (= Myriotrema concretum (Fée) Hale). The Lindig collection 
(BM and FH-Tuck.) contains psoromic acid and is correctly identified as Myriotrema 
clandestinum (Fée) Hale. While M. concretum also contains psoromic acid, there are fairly 
recognizable differences between the two species in apothecial characters, as outlined by 
Salisbury (1978 : 413). 


32. Myriotrema santessonii (Hale) Hale in Mycotaxon 11 : 135 (1980). 
(Fig. 14g) 

Thelotrema santessonii Hale in Phytologia 26: 417 (1973). Type: Ivory Coast, Man, Santesson 

10443 bis (UPS—holotype: US—isotype). 
IcOnES. Hale, 19736 : 420 fig. 8. 
Thallus greenish ashy, c. 6cm broad, shiny, with numerous hollow pustular outgrowths 
c. 1 mm high, 0°3-0°5 mm diam; cortex dense, 10-12 um, with some aculeate hyphae, 
splitting and exfoliating; algal layer continuous, 15 wm; medulla to 60 um with dense 
crystals; apothecia not seen (in type specimen immersed, 0°1-0°3 mm diam, the exciple 
fused, colourless; pore round c. 0°! mm diam; hymenium 80-90 um; spores colourless, 
muriform, 10-13 x 20-24 wm, 1-2 x 6-8 loculate, I— (apothecial characters taken from the 
type collection). 


CHEMISTRY. Stictic and constictic acids. 

HaBiTaT. On Ficus in disturbed areas at lower elevations (200 m). 

DISTRIBUTION. Ivory Coast, Tanzania, Sri Lanka, Java. 

ADDITIONAL SPECIMENS. Ratnapura District, Ratnapura, Wheeler 12471 (US). 
OBSERVATIONS. Although the single Sri Lankan specimen lacks apothecia, the chemistry and 


294 M. E. HALE 


characteristic pustules are sufficient for a tentative identification. The apothecial characters 
given here are based on the type (Hale, 19735: 417). 


33. Myriotrema subconforme (Nyl.) Hale in Mycotaxon 11 : 135 (1980). 
(Fig. 14h) 


Thelotrema subconforme Nyl. in J. Linn. Soc. Bot. 20 : 53 (1883). Type: Malaya, Malacca, May 1864, 
Maingay 64 (BM—lectotype; H-Nyl. 22587 (as no. 77), FH-Tuck., G—isolectotypes). 


Thallus pale greenish gray, 6-8 cm broad, shiny, continuous, rather fragile; cortex dense, 
10-15 wm, with aculeate hyphae, splitting into layers 5m thick; algal layer contin- 
uous, 15-18 wm; medulla 10-90 wm, with crystals; apothecia numerous, immersed to 
slightly raised, 0'2-0°3 mm diam, the exciple fused, colourless; pore round, 0:07-0'1 mm 
diam; hymenium 60-70 um; spores colourless, muriform, 5-6 x 10-18 wm, 1-3 x 4-6 locu- 
late, I+. 


CHEMISTRY. No substances present. 
HaBiTAT. Lower trunk to canopy of trees in rain forest at lower elevations (150-350 m). 
DISTRIBUTION. India, Sri Lanka, Malaya, Philippines, Java, Sarawak, Solomon Islands. 


ADDITIONAL SPECIMENS. Hale collections: 4 (50 510), 7a (50 192), 9 (46 162). Matara 
District, Dikwella, Santesson 25 744 (S). 


OBSERVATIONS. This Asian species is closest to M. album. Nylander compared it to 
Thelotrema conforme Fée (= Myriotrema conforme (Fée) Hale), a New World species with 
psoromic acid. 


34. Myriotrema terebrans (Nyl.) Hale in Mycotaxon 11 : 135 (1980). 
(Fig. 141) 


Thelotrema terebrans Nyl. in Bull. Soc. Linn. Normandie Il, 7: 166 (1873). Type: India, Andaman 
Islands, 1867, Kurz 80 (H-Nyl. 22771—lectotype; BM, UPS, ZT—isolectotypes). 
Ocellularia terebrans (Nyl.) Zahlbr., Cat. Lich. Univers. 2 : 602 (1923). 


Thallus ashy mineral gray, 4-6 cm broad, shiny, continuous; cortex very dense, 10-12 wm, 
with conspicuous aculeate hyphae, splitting into sheets internally; algal layer continuous, 
10 um; medulla to 60 wm with inclusions of cortical layers and crystals; apothecia numerous, 
immersed in the medulla, the exciple poorly developed, fused, colourless; hymenium 
45-65 um; spores colourless, transversely septate, 4-5 x 10-15 wm, 4-6 loculate, I+. 


CHEMISTRY. Stictic and constictic acids. 

HABITAT. Canopy branches in rain forest at lower elevations (150-450 m). 
DISTRIBUTION. India, Sri Lanka. 

ADDITIONAL SPECIMENS. Hale collections: 6 (47 202), 8 (47 002). 


OBSERVATIONS. Myriotrema terebrans has not been mentioned in the literature since 
Nylander’s original publication. It is in the ‘Thelotrema album’ group and can be distin- 
guished by the chemistry. On the world level there are no comparable stictic acid-containing 
species. 


35. Myriotrema terebratulum (Nyl.) Hale in Mycotaxon 11 : 135 (1980). 
(Fig. 15a) 
Thelotrema terebratulum Nyl. in Annls Sci. nat. (Bot.) V, 7 : 315 (1867). Type: Colombia, Rio Negro, 
1863, Lindig 129 (H-Nyl. 22637—lectotype; FH-Tuck., G, PC—isolectotypes). 
Ocellularia terebratula (Nyl.) Mill. Arg. in Mém. Soc. Phys. Hist. nat. Geneve 29(8) : 12 (1887). 
Ocellularia galactina Zahlbr. in Annis Cryptog. exot. 5:216 (1932). Type: South Africa, Cape 
Province, Knysna, Jan. 1928, van der Byl 673 (W—lectotype; LD—isolectotype). 
Icongs. Hale, 1974a : 26 fig. 13e (lectotype) and fig. 13i.—Hale, 1978a : 32 fig. 8d. 


THELOTREMATACEAE IN SRI LANKA 295 


Thallus ashy greenish gray, 8-14 cm broad, smooth and shiny, continuous; cortex very 
dense, about 30 wm, with abundant aculeate hyphae, splitting and exfoliating, the layers 
about 4 um thick; algal layer continuous, 15 wm; medulla 50-200 um with remnants of 
cortical layers and crystals; apothecia numerous, immersed in the medulla, c. 0°2 mm diam, 
the exciple partially free at the tips, colourless, sometimes visible through the pore; pore 
round, about 0°05 mm diam, white rimmed; hymenium 60-70 um; spores colourless, 
transversely septate, 4-7 x 9-15 wm, 34 loculate, I+. 


CHEMISTRY. Psoromic and norpsoromic acids. 


HaBITAT. Tree trunks along open trails and in canopy in rain forest at low to mid elevations 
(300-850 m). 


DISTRIBUTION. United States, West Indies, Central and South America, South Africa, India, 
Sri Lanka, Taiwan, Sabah, New Caledonia, Solomon Islands. 


ADDITIONAL SPECIMENS. Hale collections: 1 (50 222), 3a (46 310), 4 (50 493). Thwaites 
collections: C. L. 145 (BM, PDA), C. L. 174 (PDA), C. L. 175 (BM, PDA) (as ‘Lgt. 87, 
Thelotrema myriotrema’) (mixtures of Thelotrema platysporum and Myriotrema des- 
quamans in BM and UPS). 


OBSERVATIONS. This is one of the most characteristic canopy-inhabiting lichens in the 
tropics. Its relationship to M. album, M. microporum, and M. olivaceum is discussed above 
under these names. Nylander mentioned two other confusable species in his original descrip- 
tion: Thelotrema microporellum Nyl. (=Myriotrema microporellum (Nyl.) Hale), which 
contains hypoprotocetraric acid (Hale, 1978a : 24), and Thelotrema plurifarium Nyl., which 
I consider to be a synonym of M. microporellum. 

Leighton called C. L. 174 7. myriotrema Nyl. (Th. album Fée), identifying it from his set 
of Wright’s Lichenes Cubae no. 127 but not sending it on to Nylander for confirmation. The 
specimen is a mixture of M. terebratulum and Thelotrema platysporum (BM). The Cuban 
material (n. 127) contains hypoprotocetraric acid and can be determined as M. micro- 
porellum. 


36. Myriotrema thwaitesii Hale sp. nov. 
(Fig. 1 5b) 

Thallus corticola, epiphloeodes, opacus, continuus, cinereo-albus, 8-10 cm latus; apothecia 
numerosa, dispersa, immersa, 0:2-0°3 mm diametro, excipulo incolore, columella nulla; ostiolum 
rotundatum, 0:05-0:08 mm diametro, albo-cinctum; hymenium 80-90 um altum; sporae incolores, 
transversim septatae, octonae, 5-8 x 13-22 wm, 5-7 loculatae, I+ coeruleae. 

Typus: Sri Lanka, Sabaragamuwa Province, Kegalla District, Maliboda to Theberton, elev. 800 m, 
16 Mar. 1978, M. E. Hale 50 221 (US—holotypus; BM—isotypus). 

Thallus pale greenish to yellowish gray, 8-12 cm broad, dull, continuous; cortex loosely 

organized and irregularly pored, 15-18 um; algal layer continuous, 10-15 wm; medulla to 

10 um, mostly hypophloeodal; apothecia numerous, immersed in part in the periderm, 

0:2-0°3 mm diam, the exciple fused, reddish brown, columella lacking; pore round, 

0:05-0:08 mm diam, becoming white rimmed; hymenium 80-90 wm; spores colourless, 
transversely septate, 5-8 x 13-22 wm, 5-7 loculate, I+. 


CHEMISTRY. ‘Thwaitesii’ unknown (P+ orange red). 


HaBItaT. Trees along trail and on lower trunks in rain forest at low to mid elevations. 
(150-1100 m). 

DISTRIBUTION. Sri Lanka. 

ADDITIONAL SPECIMENS. Hale collections: | (50 256), 5 (46 282), 9 (46 194). Thwaites collec- 
tions: C. L. s.n. (PDA), C. L. 144 (BM, mixed with Myriotrema clandestinum) (as ‘Lgt. 76. T. 
microporum’). 


296 M. E. HALE 


‘ ae 
ar 24 oe 4 


Fig. 15 Species of Myriotrema. (a) M. terebratulum (Hale 50 222). 
(c) M. wightii (Wheeler 12 465 in US). See Fig. 71 for scale. 


ea 


(b) M. thwaitesii (Hale 50 221). 


es he 


OBSERVATIONS. The dull appearance of the thallus surface can be correlated with the poorly 
organized cortex. This trait, together with the P+ unknown, sets the species apart from others 
in the M. album complex. The substance reacts H,SO,+ dull orange and falls a little below 
the ‘praestans’ unknown from Dominica (see Hale, 1974a: 12). It may be related to 
thamnolic acid. 


37. Myriotrema wightii (Taylor) Hale in Mycotaxon 11 : 135 (1980). 
(Fig. 1 5c) 


Endocarpon wightii Taylor in Lond. J. Bot. 6: 155 (1847). Type: India, Madras, Wight s.n. 
(FH-Tayl.—lectotype; BM, G—isolectotypes). 

Thelotrema subconcretum Leighton in Trans. Linn. Soc. Lond. 27: 169 (1869). Type: Sri Lanka, 
Peradeniya, Thwaites C. L. 172 (BM—lectotype; G, NY, PC, PDA, S, UPS, W—isolectotypes). 

Leptotrema wightii (Taylor) Mill. Arg. in Flora, Jena 65 : 499 (1882). 

For additional synonymy see Sailsbury, 1971a: 5 and Hale, 1974a: 43 and 1978a: 55. 

IcoNES. Redinger, 1936: 112 fig. 71—Salisbury, 1971a: 6 fig. 1—Hale, 1978a:55 fig. 15a (errone- 
ously given as fig. 1 5b). 


Thallus pale tannish mineral gray, 4-7 cm broad, thick and almost coriaceous, shiny, 
smooth and continuous, grainy; cortex lacking, a heavily pored polysaccharide sheet 
covering the underlying tissues; medulla dense and cellular, to 600 um, organized in vertical 
blocks with algae in columns, with or without deep red crystalline masses, dense colourless 
crystals also present; apothecia immersed in the medulla, 0:2-0°3 mm diam, the exciple | 
fused, often poorly developed; pore round, 0:05—0:1 mm diam, white rimmed; hymenium 
180-200 um; spores brown, muriform, 10-14 x 14-18 wm, 2 x 4 loculate, uniseriate, I—. 


CHEMISTRY. No substances present except for K+ purple pigment. 
HABITAT. On Diospyros ebenum in shade at low elevations. 


DISTRIBUTION. United States, West Indies, Mexico, Central and South America, Portugal, 
India, Sri Lanka, Philippines, Sarawak, Australia, Hawaii. 


ADDITIONAL SPECIMENS. N. Central Province, Wilpattu National Park, Wheeler 12 463 (US). 
Thwaites collections: C. L. 123 (H-Nyl. 3888, PDA). 


OBSERVATIONS. Salisbury (1971a: 5) discussed this species fully. The red medullary crystals 
may rarely be absent. The thallus is coarse and thick. I did not collect it in the rain forests of _ 
Sri Lanka; it occurs in drier secondary forests. R 


THELOTREMATACEAE IN SRI LANKA 297 
III. OCELLULARIA G. Meyer, nom. cons. 


Nebenstund : 327 (1825). 


Ascidium Fée, Essai Crypt. : XLII (1824) (nom. gen. rejic.). 

Porophora G. Meyer, Nebenstund. : 326 (1825). 

Ectolechia Massal., Alcun Gen. Lich. : 10 (1855). 

Leptotrema Mont. & Bosch in Junghuhn, Pl. Junghuhn. 4 : 483 (1855). 
Stegobolus Mont., Syll. Gen. Sp. Crypt. : 362 (1856). 

Macropyrenium Hampe in Massal. in Atti R. Ist. veneto Sci. Ill, 5 : 329 (1860). 
Rhabdodiscus Vainio in Suomal. Tiedeakat. Toim. A, 15(6) : 184 (1921). 


Thallus crustose, epiphloeodal or rarely hypophloeodal; cortex dense, often splitting and 
exfoliating, or less commonly loosely organized, lacking in hypophloeodal species; medulla 
well developed to nearly hypophloeodal; apothecia emergent or rarely immersed, with a 
distinct thalline rim, the rim incurved to suberect, usually forming a discrete pore; exciple 
fused, carbonized; central columella often present, carbonized, simple to becoming reticu- 
late; spores colourless or brown, transversely septate or muriform, I+ blue or rarely I—. 


TYPE SPECIES. Thelotrema obturatum Ach. 
NUMBER OF SPECIES. About 200. 


OBSERVATIONS. This genus forms the most homogeneous group in the family. A few species 
in Sri Lanka lack a columella (see discussion under O. rhicnopora); the majority have one. It 
is possible to subdivide the columellate species into those with a simple columella (in part 
Salisbury’s ‘7. discolor’ group and the ‘7. cavatum’ group, these being separated by 
columella height : width ratio) and those with a reticulate columella (the ‘7. discoideum’ 
group) (Salisbury, 1978 : 407). There are, however, many transitional stages between these 
types, and I have not attempted to classify the species in this way. 


Key to the species 

Spores colourless . 

Spores brown 

Spores transversely septate. 

Spores muriform. . 

Spores large, 70-200 um long : 

Spores small to medium sized, 20-65 um: 

Central columella present . 

Central columellalacking .. to eee eee 
Thallus distinctly verruculose; no lichen substances present : ; 8. O. dolichotata 
Thallus smooth to weakly verruculose; lichen substances present. iss A Saas 
Olnvacea unKknownpresent =| . . =. =. =. - =. 26. O. nylanderiana 
Hypoprotocetraric acid present . é . ; : : : ; ; 41. O. triglyphica 
Hypoprotocetraricacidpresent. . . . . . . 25. O. neopertusariiformis 
‘Pertusariiformis’ unknown present. . . . . . . 30. O. pertusariiformis 
Medulla wholly orinpartorange . . . . . . . . 33. O. punctulata 
Medulla white . . eee Aer oS ee > ee 9 
Pore rim and tip of columella pale orange : ae 10 
Pore rim and tip of columella white, blackening, or ¢ concolourous with the thallus : ? 11 
Medulla P+ yellow (psoromic acid present) . 7 : é é 4 15. O. lankaensis 
Medulla P—(‘chonestoma’ unknown present) . : : 6. O. croceopora 
Thallus dull, lacking a cortex; pore area usually darkening . eae 34. O. pyrenuloides 
Thallus shiny, corticate; pore areaconcolourousorwhite . . . . . . . IP 
2 Central columella present . ; 

Central columella lacking . ae 
No lichen substances present on t.l.c.; . medulla Pe. ’ 14 
Lichen substances always present; medulla P+ orange or red Pm only i in O. chonestoma) 16 
Apothecia flush to barely raised, 0°4—0°7 mm diam ‘ : ; ; 28. O. papillata 
Apothecia emergent, 0°8-1'2 mmdiam . : : : : : : Z : ; 5 15 
SOLES OS mmUIOne. 2. 2 a 6 on we RTE ets 27. O. orthomastia 


Ww 


1Nl = 
tN 
NUMNohHRWAY 


a 


ce 


OS Be ee |! ie 


oS 


— 
ay 


M. E. HALE 


Spores 18-30 um long ’ : : : , : 2. O. ascidioidea 

Medulla P—(‘chonestoma’ unknown present) ee a Oe» |e 4. O. chonestoma 

Medulla P+ orange or red F 

Pore gaping, to 0°7 mm wide; thalline rim Sunoreer columella foe becoming reticulate 
9. O. emersa 

Pore smaller, 0: 05-0: a mm diam: thalline rim incurved: columella narrow, simple 


Apothecia large andemergent, 1-1:3 mmdiam . ’ , 15. O. lankaensis 
Apothecia smaller, less than 1 mm wide, not strongly emergent . 5 j : : 3 
Pore coarsely white rimmed; ‘cinchonarum’ unknown present .. 24. O. neocavata 
Pore not strongly white rimmed; ‘cinchonarum’ unknown lacking J pide ne 
Medulla P+ yellow(psoromicacid present) . . . . . . 31. O. pluripora 
Medulla P+ orange or red (psoromic acid lacking) . 0! 
Stiticacid presentont.l.c.. .  . : ‘ ea ; j : 37. O. sticticans 
Stictic acid lacking a | e. SA ee Se 
Protocetraric acid present o1 ont. 1; Ce fre 4 ee ee 29 O. perforata 
‘Diacida’ unknown present ont.l.c. . : ; 7. O. diacida 


Pore area concolourous with the thallus; hypoprotocetraric acid present on t.l.c. 

13. O. kanneliyensis 
Pore area brownish, wrinkled; nosubstances present . . . . 36. O. rhicnopora 
Spores large, 100-360 um long . 
Spores small, less than 30 wm long 3 : : : ; 2 ; : : 
Central columella present . bd ¢ : : A, s ; 4 19. O. meer 
Central columella lacking . 
Apothecia somewhat raised to emergent, 0: 6-1- 0 mm diam: pore small, 0: 1 mm “diam 

14. O. keralensis 

Apothecia emergent to strongly emergent, |- I: 3mm diam: pore larger, to 0-2 mm diam 
Spores turning brownish at maturity; no lichen substances present 23. O. monosporoides 
Spores colourless at maturity; hypoprotocetraric acid present . : 10. O. eumorpha 
Central columella lacking; disc orange yellow pruinose ee Dn fe: 3. O. aurata 
Central columella present; disc not visible or white pruinose 
Thallus dull, whitish gray, cortex lacking 
Thallus shiny, whitish green; cortex present 


Columella much wider than high; pore 0:2-0°3 mm diam les 16. O. leucomelaena 
Columella narrow, higher than wide; pore less than 0-1! mm diam t : By? 

Thallus P+ orange (stictic acid present) . 4 ; ; d : 22. O. melanotremata 
Thallus P—(no substances present) . : : 39. O. tenuis 
Columella reticulate at maturity; pore gaping, 0: 2-0: 4r mm diam , 32. O. polillensis 
Columella simple; pore smaller, 0: 1-0-2 mm diam : A i : : : 

Protocetraric acid present ont.l.c. . : : 38. O.subsimilis 


Fumarprotocetraric acid present on t.l.c. (with eee: Plprolecsnene acid) 

40. O. thelotremoides 
Apothecia elongate at maturity . : 5 : : : : ; . 17. O. lirelliformis 
Apothecia round . i 
Thallus dull, whitish gray; cortex lacking. 
Thallus shiny, whitish green; cortex present Sy arth, Vey ire lg 
Columella narrow, 60-180 um diam; apotheciaraised. . . 21. O. meiospermoides 
Columella broader, to 300 um diam, becoming reticulate; apothecia flush a) ae 
: : F . ? ; : ; 5 ; , : : : ; ’ 20. O. meiosperma 
Apothecia deeply immersed in periderm; pore very tiny, to 0:5 mm diam. 11. O. exuta 
Apothecia not deeply immersed, raised at maturity; pore 0°1 mm diam, white rimmed 
; : : ; : ; : : } ‘ : ‘ : : ; : 35. O. rassagala 
Spores large, 120-210 wm long . sn Os: : : p : : 23. O. monosporoides 
Spores small, 10-30 wm long 3 : 3 ; 3 tue te 
Columella simple 
Columella becoming reticulate at maturity 


Thallus verruculose and white spotted; spores 22-30 um . fig O. albomaculata 
Thallus smooth, not spotted; spores 12-20 um : , : . 18. O. marivelensis 
Thalline rim suberect, often jagged . : ; : : eae ; 3 5. O. crassa 


Thalline rim more or less incurved, entire : : : : ; : 12. O. fissa 


25 
28 


26 


27 


29 
30 
32 


31 


33 


35 
36 
30 


39 
40 
4] 


THELOTREMATACEAE IN SRI LANKA 


—_ dl 
b) O. ascidioidea (Hale 47 035). 


croceopora (Hale 46 129). (g) O. diacida (Hale 50 258). (h) O. dolichotata (Hale 46 210). (i) O. 
emersa (Hale 50 145). See Fig. 71 for scale. 


299 


300 M. E. HALE 


1. Ocellularia albomaculata Hale sp. nov. 
(Fig. 16a) 


Thallus corticola, epiphloeodes, verruculosus, continuus vel fissurinus, nitidus, viridi-albus, ad 10 cm 
latus; apothecia emergentia, 0°8-1-2 mm diametro, amphithecio verrucoso, excipulo fuligineo, 
columella 250-300 um diametro; ostiolum rotundatum, 0°2-0°-4mm diametro, albo-cinctum; 
hymenium 150-180 4m altum; sporae fuscae, murales, 9-10 x 22-30 um, 1-2 x4 loculatae, 
uniseriatae, I+ coeruleae. 

Typus: Sri Lanka, Southern Province, Kanneliya Forest Reserve southeast of Hiniduma, 16 Feb. 
1976, M. E. Hale 46 140 (US—holotype; AMH, PDA—isolectotypes). 


Thallus pale greenish gray, up to 10cm broad, finely verrucose, shiny, continuous to 
fissured; cortex dense, 12-15 wm, splitting with some exfoliation; algal layer 15-18 um, more 
or less continuous between crystals; medulla 80-150 um, with large crystal masses to 100 um 
wide; apothecia emergent, 0°8-1-2 mm diam, the amphithecium verruculose, exciple 
carbonized; columella present, 250-300 um broad; hymenium 150-180 wm; spores brown, 
muriform, 9-10 x 22-30 wm, 1-2 x 4 loculate, uniseriate, I+. 


CHEMISTRY. No substances present. 

HasiTatT. Mid bole to canopy of trees in rain forest at low elevations (150 m). 
DISTRIBUTION. Sri Lanka. 

ADDITIONAL SPECIMENS. Hale collections: 9 (46 148). 


OBSERVATIONS. This species is characterized by the greenish, shiny, verruculose, grainy and 
spotted cortex and the gaping pore through which the top of the columella 1s visible. 


2. Ocellularia ascidioidea Hale sp. nov. 
(Fig. 16b) 


Thallus corticola, epiphloeodes, continuus, aetate rimosus, cinereo-albus, 6-10 cm latus; apothecia 
numerosa, solitaria, emergentia, 0°8—1‘0 mm diametro, epithecio fuligineo, columella 150-190 um 
diametro, fusca; ostiolum rotundatum, 0°:2-0°3 mm diametro; hymenium 150-190 um altum; sporae 
octonae, incolores, transversim septatae, 5—8 x 18-30 wm, 6-8 loculatae, I+ coeruleae. 

Typus: Sri Lanka, Southern Province, Matara District, Morawaka Hill southwest of Deniyaya, 15 
Feb. 1976, M. E. Hale 47 035 (US—holotypus; BM—isotypus). 

Thallus whitish mineral gray, 6-10 cm diam, shiny, continuous; cortex dense, 8-10 um; 

algal layer 10-15 wm; medulla variable, 10-20 um, mostly hypophloeodal; apothecia 

numerous, moderately emergent, 0°8-1-0 mm diam, the exciple carbonized; columella 
distinct, 150-190 wm diam; pore 0:2-0°3 mm diam, often filled with tip of columella; 

hymenium 150-190 wm; spores colourless, transversely septate, 5-8 x 18-30 um, 6-8 

loculate, I+. 


CHEMISTRY. No substances present. 

HABITAT. Lower bole, rarely into the canopy, in rain forest at 150-600 m elevation. 
DISTRIBUTION. Sri Lanka. 

ADDITIONAL SPECIMENS. Hale collections: 4 (47 082), 7b (51 066), 8 (47 012), 9 (46 175). 


OBSERVATIONS. This species has clearly emergent apothecia and a well developed columella. 
The most closely related species, O. papillata, has flush to barely raised apothecia and a 
variable, often weakly developed columella. 


3. Ocellularia aurata (Tuck.) Hale in Mycotaxon 11 : 136 (1980). 
(Fig. 16c) 


Thelotrema auratum Tuck. in Proc. Am. Acad. Arts Sci. 5: 408 (1862). Type: Cuba, Wright 133 
(FH-Tuck.—lectotype; BM, G, L, PC, UPS—isolectotypes). 


THELOTREMATACEAE IN SRI LANKA 301 


For additional synonymy see Hale (1978a : 40). 
IcongEs. Hale 1978a: 5 fig. 1f (lectotype). 


Thallus greenish to whitish gray, 5-10 cm broad, shiny, fissured; cortex dense, variably 
thickened, 18-30 wm; algal layer continuous, 15 wm; medulla to 90 um with crystals; 
apothecia conspicuous, strongly emergent, 0°8-1:2 mm diam, the exciple carbonized; 
columella lacking; pore gaping, 0°2-0°4 mm diam, disc and inner walls pale yellowish to 
orange-red pruinose; hymenium 100-110 ym; spores colourless, muriform, 6-8 x 12- 
18um, 1-2 x 4-5 loculate, I+. 


CHEMISTRY. Protocetraric acid and unidentified K+ purple pigments. 
HABITAT. Canopy branches in rain forest at lower elevations (150-350 m). 


DISTRIBUTION. West Indies, Central and South America, Sri Lanka, Philippines, Sarawak, 
New Caledonia, Samoa, Solomon Islands. 


ADDITIONAL SPECIMENS. Hale collections: 3a (46 308), 4 (47 124, 50 521), 9 (46 212). 


OBSERVATIONS. This pantropical species generally occurs in the canopy in rain forest at low 
elevations. The pigmented disc is easily seen through the large pore. 


4. Ocellularia chonestoma (Leighton) Zahlbr., Cat. Lich. Univers. 2 : 586 (1923). 
(Fig. 16d) 


Thelotrema chonestomum Leighton in Trans. Linn. Soc. Lond. 27 : 171 (1870). Type: Sri Lanka, south 
ofisland, Thwaites 168 (BM—lectotype; G, PC, PDA, UPS, W, isolectotypes).. 


Thallus ashy whitish to tannish mineral gray, 8-15 cm broad, shiny, continuous; cortex 
weakly developed and irregularly pored, up to 8 wm thick with some aculeate hyphae; algal 
layer 10 um; medulla mostly hypophloeodal, embedded between periderm layers; apothecia 
numerous and conspicuous, strongly emergent and in part basally constricted, 0°8—1°5 mm 
diam, the amphithecium with a thick medulla easily abraded and exposed, exciple 
carbonized; columella well developed, 90-300 um diam; pore round, 0°2-0°5 mm diam; 
hymenium 90-140 um; spores colourless, transversely septate, 5-6 x 18-30 um, 6-8 
loculate, I+. 


CHEMISTRY. ‘Chonestoma’ unknown and the upper ‘olivacea’ unknown (P—). 


Hasitat. Base, lower bole, saplings, rarely into the canopy in rain forest at lower elevations 
(150-800 m). 


DISTRIBUTION. Sri Lanka, Philippines, Sarawak, Sabah, Solomon Islands. 


ADDITIONAL SPECIMENS. Hale collections: | (50 231, 50 233, 50, 235, 50 257), 3a (46 309, 
46 313, 46 324, 46 344, 46 378), 4 (47 109, 50 399, 50 415, 50 417, 50 421, 50 494, 50 522), 6 
(47 079, 47 089, 47 108, 47 149, 47 157, 47 164, 47 167, 47 174, 47 185, 47 216), 7a (SO 179, 
50 184), 7b (50 299, 50 311, 51 001), 9 (46 133, 46 157, 46 176, 46 193, 46 197, 46 204, 
46 206, 46 219, 46 394). Thwaites collections: C. L. 269 (PDA), C. L. s.n. (BM PDA (as ‘Lgt. 
101, Ascidium chonestomum), C. L. 169 (BM, PDA, UPS) (as ‘Let. 88, 7. cavatum var. 
confertum’). 


OBSERVATIONS. This is the most common member of the family in Sri Lanka. Although it 
occurs, much less commonly, in dipterocarp forest as far east as the Solomon Islands, there 
are no records from neighbouring India. One character not mentioned by Leighton is the 
tendency for the emergent apothecia to become abraded, revealing the thick, pure white 
medulla of the amphithecium and contrasting strongly with the tannish green thallus. 

The ‘chonestoma’ unknown is a H,SO,+ gray spot above the ‘olivacea’ series in hexane 
system. These two P— compounds are probably closely related depsides or depsidones. 

Leighton’s determination of ‘7. cavatum Nyl., var. confertum Nyl.’ proved to be O. 
chonestoma. The name, not checked by Nylander, appears to have been taken from a 


302 M. E. HALE 


comparison of the Sri Lankan material with Lindig’s Colombia collection so named by 
Nylander. This taxon, Ocellularia conferta (Nyl.) Dodge, is a psoromic acid-containing 
member of Myriotrema and may be a good species. 


5. Ocellularia crassa (Miill Arg.) Hale in Mycotaxon 11 : 136 (1980). 
(Fig. 16e) 


Leptotrema crassum Mill. Arg. in Flora, Jena 65:332 (1882). Type: Java, Junghuhn 158 
(L—lectotype; H-Nyl., G—isolectotypes). 

Leptotrema fallax Mill Arg. in Flora, Jena 70:62 (1887). Type: Australia, Richmond River, 
Hodgkinson s.n. (G—lectotype). 

Leptotrema integrum Mill. Arg. in Flora, Jena 70 : 399 (1887). Type: Australia, Queensland, Sayer 
s.n. (G—lectotype). 

Thelotrema vesiculiferum Vainio in Suomal. Tiedeakat. Toim. A, 15(6):175 (1921). Type: 
Philippines, Prov. Sorsogon, Irosin, Dec. 1915, Elmer 1406 (TUR—lectotype; FH, US, 
W—isolectotypes). 

Leptotrema vesiculiferum (Vainio) Zahlbr., Cat. Lich. Univers. 2 : 641 (1923). 


Thallus greenish mineral gray, 4-12 cm broad, shiny, continuous; cortex dense, 12-15 um, 
with some aculeate hyphae; algal layer continuous, 15-20 wm; medulla 10-100 um, with 
dense crystals; apothecia semi-emergent, the thalline wall becoming suberect, the rim jagged, 
exciple carbonized; columella entire to reticulate, 140-250 um diam, the tip white pruinose; 
pore gaping, 0°2-0°-4 mm diam; hymenium 65-90 wm; spores brown, muriform, 5-10 x 
9-15 wm, 1-2 x 3-5 loculate, I—. 


CHEMISTRY. Psoromic acid with or without norpsoromic acid. 
HABITAT. Base to mid trunk of trees in rain forest at low to mid elevations (150-850 m). 
DISTRIBUTION. Sri Lanka, Philippines, Java, Sabah, Australia. 


ADDITIONAL SPECIMENS. Hale collections: 1 (50 208, 50 236), 2 (51 103, 51 117, 51 125, 
51 131,51 155,51 160,51 190, 51 210), 3a (46 295), 3b (46 386), 4 (47 074, 47 076, 47 104, 
47 215, 50 429), 5 (46 275). 


OBSERVATIONS. As in other genera in the family, the species with psoromic acid tend to form 
intergrading aggregates. The large synonymy I propose reflects the problems in defining 
columellar development, degree of emergence of the thalline wall, and variation in pore 
width. I have here defined O. crassa as having a greenish, shiny thallus, relatively thick 
cortex (although the thallus appears thin under a hand lens), jagged, suberect thalline rim, 
and gaping pore. It occurs at lower elevations. 

Ocellularia crassa has frequently been compared (and confused) with O. fissa (see below). 
Miller Argoviensis considered O. crassa to be near O. fissa, although I doubt that he saw the 
Nylander type. Externally O. fissa appears to have a thicker thallus, even though it is not 
much, if at all, thicker when measured against O. crassa. It also has a more whitish colour 
and less emergent apothecia with a smooth rim. Finally it seems to occur at higher 
elevations, above the range of O. crassa. 


6. Ocellularia croceopora Hale sp. nov. 
(Fig. 1 6f) 


Thallus corticola, epiphloeodes, continuus vel fissus, nitidus, pallide viridi-albus, 8-15 cm latus; 
apothecia numerosa, prominentia, emergentia, solitaria vel rare 2: nae aggregata, 0°9-1-2 mm 
diametro, columella 180-230 um; ostiolum rotundatum, 0°2-0°'4 mm diametro, profundum, vix 
albo-annulatum, margine et apice columellae pallide croceo; hymenium 65-120 um altum; sporae 
incolores, tranversim septatae, octonae, 3-5 x 9-15 wm, 4 loculatae, I+ coeruleae. 

Typus: Sri Lanka, Southern Province, Galle District, Kanneliya Forest Reserve, elev. 150 m, 16 
Feb. 1976, M. E. Hale 46 129 (US—holotypus; BM—isotypus). 


Thallus greenish to whitish mineral gray, 8-15 cm broad, shiny; cortex dense, 12-18 wm, 


THELOTREMATACEAE IN SRI LANKA 303 


with aculeate hyphae, splitting into layers internally; algal layer continuous, 15-20 um; 
medulla to 10 wm, mostly hypophloeodal; apothecia common, emergent, 0°9-1:2 mm diam, 
the exciple carbonized, columella 180-230 um; pore round, 0:2-0°4 mm diam, columella 
surface and inner wall pale orange; hymenium 65-120 um; spores colourless, transversely 
septate, 3-5 x 9-15 wm, 4-5 loculate, I+. 


CHEMISTRY. ‘Chonestoma’ unknowns (P—). 


HaBiTaAT. Base and lower trunk, rarely into the canopy, in rain forest at lower elevations 
(150-200 m). 


DISTRIBUTION. Sri Lanka. 


ADDITIONAL SPECIMENS. Hale collections: 9 (46 151, 46 169, 46 181, 46 186, 46 201, 46 209, 
46 214). 


OBSERVATIONS. Ocellularia croceopora is very similar to O. lankaensis externally, with a 
pale orange-pruinose pore area but smaller spores and different chemistry (O. lankaensis has 
psoromic acid). Curiously the species was collected frequently at Hiniduma in the Kanneliya 
Forest Reserve but at no other localities. At the same time O. /ankaensis was represented at 
most of the low elevation localities but only once at Hiniduma. 


7. Ocellularia diacida Hale in Mycotaxon7 : 378 (1978). 
(Fig. 16g) 


Type: India, Karnataka, Devimane Ghat, evergreen forest, 26 Feb. 1977, M. E. Hale 47933 (US— 
holotype; AMH—isotype). 
Icons. Hale, 19785 : 384 fig. 2. 


Thallus greenish to dull brownish gray, to 10 cm broad, dull, smooth to fissured with age; 
cortex dense, 8-10 um; algal layer continuous, 15-18 um; medulla up to 20 wm thick, mostly 
hypophloeodal; apothecia numerous, nearly flush to semi-emergent, 0°3-0°4 mm diam, the 
exciple fused, reddish brown; columella 45-50 wm diam; pore round, 0°! mm diam, white 
rimmed; hymenium 110-130 um; spores colourless, transversely septate, 5-6 x 15-30 um, 
7-8 loculate, I+. 


CHEMISTRY. Unknown substances (‘diacida’ unknowns, P—). 
Hasitat. Tree trunks along trail at mid elevations (800 m). 
DISTRIBUTION. India, Sri Lanka. 

ADDITIONAL SPECIMENS. Hale collections: | (50 258). 


OBSERVATIONS. The single specimen from Sri Lanka is identical with the Indian population. 
I can add nothing new about the identity of the unknown spot, a H,SO,+ brown double spot 
near norstictic. The species is probably related to O. perforata, which contains protocetraric 
acid. 


8. Ocellularia dolichotata (Nyl.) Zahlbr., Cat. Lich. Univers. 2 : 589 (1923). 
(Fig. 16h) 


Thelotrema dolichotatum Nyl., Sert. Lich. Trop. Labuan Singapore : 19 (1891). Type: Singapore, Nov. 
1879, Almquist s.n. (H-Nyl. 22748—lectotype; S—isolectotype). 

Thelotrema siamense Vainio in Hedwigia 47:175 (1907). Type: Thailand, Ko Chang, Klong 
Sarlakpet, 1900, Schmidt XXX VII (TUR—lectotype; C—isolectotype). 

Ocellularia siamensis (Vanio) Zahlbr., Cat. Lich. Univers. 2 : 600 (1923). 


Thallus ashy white, 8-12 cm broad, shiny, verruculose and grainy; cortex cellular, 10-20 um 
thick, with some aculeate hyphae; algal layer 10 wm; medulla variable, to 10 wm, with 
numerous crystals, largely hypophloeodal; apothecia rather sparse, emergent, densely ver- 
ruculose, 1-1-1-4 mm diam, the exciple carbonized; columella 150-300 um diam; pore 


304 M. E. HALE 


round, 0°1-0:2 mm diam; hymenium 200-250 um; spores colourless, transversely septate, 
15-20 x 120-200 wm, 12-15 loculate, 2/ascus, I+. 


CHEMISTRY. No substances present. 


HABITAT. Base to mid bole, rarely into the canopy, in rain forest at low elevations 
(100-300 m). 


DISTRIBUTION. Sri Lanka, Thailand, Peninsular Malaysia, Philippines, Sabah, Sarawak, 
Solomon Islands. 


ADDITIONAL SPECIMENS. Hale collections: 3a (46 296), 4 (47 098), 9 (46 146, 46 161, 46 183, 
46 196, 46 210). 


OBSERVATIONS. This Asian species is characterized by the large apothecia and strongly 
verruculose thallus, or as Nylander described it, ‘minute granulato-rugatulus’. Vainio differ- 
entiated his Thelotrema siamense from T. dolichotatum by the presence of a columella and 
asci with 2-3 spores. In fact Nylander’s type also has a columella, and while the spores in T. 
siamense are reported to be 12-14 x 30-110 wm, a measurement which I confirmed, they 
appear to be immature; the hymenium is infested with a parasitic fungus. 


9. Ocellularia emersa (Krempelh.) Mull. Arg. in Flora, Jena 69 : 310 (1886). 
(Fig. 161) 


Thelotrema emersum Krempelh. in Flora, Jena 59: 221 (1876). Type: Brazil, Glaziou 3189 (M— 
lectotype; C, G, H, UPS, W—isolectotypes). 

Thelotrema efformatum Krempelh. in Flora, Jena 59 : 221 (1876). Type: Brazil, Glaziou 3190 (M— 
lectotype; C, G, H, UPS, W—isolectotypes). 

Thelotrema schizostomum Krempelh. in Flora, Jena 59 : 222 (1876) (non Tuck. ex anno 1862). Type: 
Brazil, Glaziou 3260 (M—lectotype; C, H, UPS, W—isolectotypes). 

Ocellularia efformata (Krempelh.) Miill. Arg. in J. Linn. Soc. (Bot.) 30 : 452 (1895). 

Ocellularia emersella Mill. Arg. in J. Linn. Soc. (Bot.) 30:453 (1895) (Based on Thelotrema 
schizostomum). 

Phaeotrema emersum (Krempelh.) Zahlbr. in Denkschr. Akad. Wiss. Wien (Math.-nat. KI.) 83 : 119 
(1909). 

ICONES. Redinger, 1936 : 40, 41 figs 17, 18. 


Thallus pale whitish green to yellowish gray, 6-10 cm broad, shiny, continuous; cortex 
dense, 8-10 wm, with aculeate hyphae, splitting and exfoliating; algal layer continuous, 
15 um; medulla 15-30 wm, with large crystals; apothecia common, semi-emergent with a 
suberect, rarely somewhat recurved thalline rim, round to irregular, 0°-S—1-0 mm diam, the 
exciple carbonized; columella variable, in part reticulated, 60-200 wm diam, apically 
pruinose; pore variable, 0:1-0°7 mm diam; hymenium 60-75 um; spores colourless, trans- 
versely septate, 4-5 x 9-12 wm, 5-7 loculate, I+. 


CHEMISTRY. Psoromic and norpsoromic acids. 
HABITAT. Lower trunks of trees in rain forest at lower elevations (150-300 m). 
DISTRIBUTION. West Indies, Central and South America, Sri Lanka. 


ADDITIONAL SPECIMENS. Hale collections: 4 (47 143), 7a (50 148), 7b (51 022, 51 047, 
51 068), 9 (46 191, 46 200). Thwaites collections: C. L. 166 (BM, PDA), C. L. 269 (PDA) (as 
‘Let. 78, Thelotrema platystomum’, mixed with Ocellularia fissa). 


OBSERVATIONS. Redinger (1936 : 39) has discussed this species rather thoroughly, consider- 
ing Thelotrema efformatum to be a synonym since the alleged spore difference—smaller 
spores in T. efformatum, according to Krempelhuber—did not hold up on re-examination. 
Ocellularia emersa is indeed variable in terms of development of the thalline rim and degree 
of reticulation of the columella, and we may be dealing with more than one species on the 


THELOTREMATACEAE IN SRI LANKA 305 


world level. Zahlbruckner (1909 : 119) was correct in seeing a relation between this species 
and Phaeotrema virens Mill. Arg. (= Ocellularia virens (Mill. Arg.) Hale). 

The Thwaites material was mixed, part being Ocellularia emersa, part O. fissa. Leighton 
had identified it as ‘7. platystomum Mont.’ without the benefit of Nylander’s advice. An 
isotype of O. platystomum (Mont.) Zahlbr. in H-Nyl. (Leprieur 491], French Guyana), how- 
ever, lacks any chemistry, has a non-reticulated columella, and is apparently hypophloeodal. 
It cannot be characterized beyond this until better collections are seen. 


10. Ocellularia eumorpha (Stirton) Hale in Mycotaxon 11 : 136 (1980). 
(Fig. 17a) 


Thelotrema eumorphum Stirton in Proc. phil. Soc. Glasg. 10: 158 (1877). Type: India, Kumaon, 
Dictrie s.n. (BM—lectotype). 


Thallus pale yellowish or straw coloured, 5—12 cm broad, shiny, smooth to rugulose; cortex 
loosely cellular, 15-20 um; algal layer continuous, 10-15 wm; medulla 0-30 wm with small 
crystals, mostly hypophloeodal; apothecia emergent to strongly emergent, solitary to loosely 
clumped in twos or threes, 1—-1°3 mm diam, the exciple carbonized; columella lacking; pore 
round, 0°1-0°'13mm diam, white rimmed and sometimes depressed; hymenium 
280-300 um; spores colourless, muriform, 25-35 x 150-180 wm, with numerous locules, 
1—2/ascus, I+. 


CHEMISTRY. Hypoprotocetraric and 4-O-demethylnotatic acids (see Culberson & Hale, 
1973). 


HaBiTAT. Lower trunks in rain forest at low to mid elevations or higher (150-2100 m). 
DISTRIBUTION. India, Sri Lanka. 
ADDITIONAL SPECIMENS. Hale collections: 2 (51 224), 5 (46 253), 9 (46 217), 10 (50 282). 


OBSERVATIONS. Ocellularia eumorpha has large emergent apothecia and externally resem- 
bles O. domingensis (see below under O. rhicnopora). A related Asian species, O. arecae 
(Vainio) Hale, differs only in having a columella, a character that has not proved to be 
especially strong in other species. Stirton compared his species with Thelotrema occlusum 
Nyl., a New World species which coincidentally has hypoprotocetraric acid but differs 
strongly in having immersed apothecia and periphysoids. Salisbury (1971b: 277) listed 
Thelotrema (Ocellularia) arecae as a synonym of T. interpositum Nyl. (= Ocellularia inter- 
posita (Nyl.) Hale). However, O. interposita is a New World species with psoromic acid and 
a strongly verruculose thallus (Hale, 1974a : 35). Another synonym he gives, 7. sanfordi- 
anum Zahlbr. (= Ocellularia sanfordiana (Zahlbr.) Hale), lacks any lichen substances and has 
a blackened pore area; it occurs only in the southern United States. A third synonym, T. 
turgidulum Vainio (Salisbury, 1971b:278), may well be correct but I have not 
yet seen the type specimen. 


11. Ocellularia exuta Hale sp. nov. 
(Fig. 17b) 


Thallus corticola, epiphloeodes, fragilis, continuus, 6—10cm latus; apothecia profunde immersa, 
0:2-0°3 mm diametro, excipulo fuligineo, columella 60 um diametro; ostiolum rotundatum, 
0:01-0:05 mm diametro; hymenium 150-160 um altum; sporae fuscae, transversim septatae, 
octonae, 6-8 x 12-24 wm, 4-6 loculatae, I—. 

Typus: Sri Lanka, Southern Province, Galle District, Kanneliya Forest Reserve, elev. 150 m, 16 
Feb. 1976, M. E. Hale 46 158 (US—holotypus; BM, PDA—isotyp)). 


Thallus whitish mineral gray, 6-10 cm broad, growing on rotted bark and easily disinte- 
grating, shiny, continuous; cortex dense but very thin, 5 wm, algal layer continuous, 15 wm; 
medulla 10-20 um with crystals, mostly hypophloeodal; apothecia inconspicuous, deeply 
immersed in the periderm, 0°2-0°3 mm diam, the exciple carbonized; columella c. 60 um 


306 M. E. HALE 


iB iis. Pes aw 3 i OR er eS 
Fig. 17. Species of Ocellularia. (a) O. eamorpha (Hale 51 224). (b) O. exuta (Hale 46 158). (c) O. 
fissa (Hale 50 283). (d) O. kanneliyensis (Hale 46 213). (e) O. keralensis (Nagarkar & Gole 


76 542 in US). (f) O. keralensis (Hale 46 281). (g) O. lankaensis (Hale 47 184). (h) O. leuco- 
melaena (Hale 50 294). (i) O. lirelliformis (Hale 51 092). See Fig. 7i for scale. 


THELOTREMATACEAE IN SRI LANKA 307 


diam; pore round, 0:01-0:05 mm diam, narrowly white rimmed; hymenium 150-160 wm; 
spores brown, transversely septate, 6-8 x 18-24 wm, 4-6 loculate, uniseriate, I—. 


CHEMISTRY. Psoromic acid. 

HaBiTAT. Canopy branches in rain forest at lower elevations (300 m). 
DISTRIBUTION. Sri Lanka. 

ADDITIONAL SPECIMENS. Hale collections: 9 (46 136). 


OBSERVATIONS. This unique species has immersed apothecia with a tiny pore and a fragile 
thallus growing on crumbling bark. It occupies the same habitat as Thelotrema platysporum 
in the rain forest. There are no related species. 


12. Ocellularia fissa (Nyl.) Hale in Mycotaxon 11 : 136 (1980). 
(Fig. 17c) 


Thelotrema fissum Nyl. in Annls Sci. nat. (Bot.) IV, 11: 258 (1859). Type: Bourbon [= Réunion], 
Richard s.n. (H-Nyl. 229659—lectotype; G, PC—isolectotypes). 

Trypethelium schizostomum Leighton in Trans. Linn. Soc. Lond. 27: 184 (1870). Type: Sri Lanka, 
Central Province, Thwaites C. L. 121 (BM—lectotype; H-Nyl. 7562, G, PDA, S—isolectotypes) 
(duplicates as Lgt. 192 in BM, S= Myriotrema fluorescens). 

Leptotrema fissum (Nyl.) Mill. Arg. in Flora, Jena 65 : 333 (1882). 

Leptotrema sandwicense Zahlbr. in Annls mycol. 10 : 372 (1912). Type: Hawaii, Oahu, Mt. Punaluu, 
Rock 60 (W—lectotype). 

Rhabdodiscus fissus (Ny1.) Vainio in Suomal. Tiedeakat. Toim. A, 15(6): 184 (1921). 

Thelotrema schizostomoides Zahlbr., Cat. Lich. Univers. 2: 628 (1923). Based on Trypethelium 
schizostomum [Non Thelotrema schizostomum Tuck. ex anno 1862, nec Krempelh. ex anno 1876.] 
Leptotrema oleosum Zahlbr. in Bot. Mag., Tokyo 61:317 (1927). Type: Japan, Yakushima, July 

1906, Faurie 2415 (W—lectotype; K YO—isolectotype). 

IcongES. Hale, 1974a : 41 fig. 19a (lectotype) and fig. 19c. 


Thallus whitish or tannish gray, 6-10 cm broad, shiny, continuous to fissured with age; 
cortex dense but rather thin, 8-10 wm, with aculeate hyphae; algal layer continuous, 15 wm; 
medulla 20-60 um, with crystals; apothecia semi-emergent, 0°7-1-1 mm diam, the thalline 
rim becoming suberect, entire, exciple carbonized; columella more or less reticulate, 
60-100 wm diam, the pruinose tip often filling the pore; pore gaping at maturity, 
0:2-0'6 mm diam; hymenium 90-120 um; spores brown, muriform, 6-7 x 8-15 um, 
1-2 x 34 loculate, I—. 


CHEMISTRY. Psoromic acid with or without norpsoromic acid. 
HaBiTAT. Tree trunks in open forest at higher elevations (2100-2200 m). 


DISTRIBUTION. West Indies, Mascarene Islands, Sri Lanka, Japan, Solomon Islands, Hawaii. 


ADDITIONAL SPECIMENS. Hale collections: 10 (50 275, 50 283), 11 (50 325, 50 327), 12 
(50 362). Thwaites collections: C. L. 166 (BM, PDA, S, UPS) (as ‘Lgt. 78. Thelotrema platy- 
stomum’) (mixture with Ocellularia emersa). Almquist collections: Pidurutalagala, A/mquist 
s.n. (H-Nyl., S) (as ‘Thelotrema epitrypum’). 


OBSERVATIONS. Nylander presented the basic features of this species in his original descrip- 
tion: ‘margo thallinus firmus erectus vel nonnihil connivens’. Zahlbruckner emphasized the 
thick thallus, although as I pointed out above it is not really much thicker than that of 0. 
crassa, a confusable relative at lower elevations in Sri Lanka. As with O. crassa, this 
psoromic acid-containing species is extremely variable and I cannot claim that the treatment 
presented here is final. 

Nylander identified the Almquist collection as ‘Thelotrema epitrypum Nyl. N. Gran. 
p. 49°. That species, which incidentally also contains psoromic acid (Lindig 2769, isolecto- 


308 M. E. HALE 


type in FH-Tuck.), has small, emergent apothecia (0°-4-0°6 mm diam) with a simple, narrow 
columella. It is not at all related to O. fissa. 


13. Ocellularia kanneliyensis Hale sp. nov. 
(Fig. 17d) 


Thallus corticola, epiphloeodes, verruculosus granulosusque, tenuis, pallide olivaceus, 8—12 cm latus; 
apothecia dispersa, separata, valde emergentia, basin constricta, 1‘2-1-5 mm diametro, excipulo 
apice fuligineo, columella nulla; ostiolum rotundatum, 0°3-0°5 mm diametro, nigro-cinctum; 
hymenium 140-150 wm altum; sporae incolores, transversim septatae, octonae, 6-7 x 24-33 um, 
6-8 loculatae, I+ coeruleae. 

Typus: Sri Lanka, Southern Province, Galle district, Kanneliya Forest Reserve, elev. 150 m, 16 
Feb. 1976, M. E. Hale 46 213 (US—holotypus). 


Thallus pale brownish to olive gray, 8-12 cm broad, densely verruculose to granular, 
continuous; cortex very dense, 15-20 um; algal layer 10 wm; medulla about 10 wm, bulging 
out with crystals inclusions in the verrucae, mostly hypophloeodal; apothecia sparse, 
strongly emergent, 1-2-1°5 mm diam, the exciple carbonized; columella lacking; pore round 
to irregular, 0°3-0°5 mm diam; hymenium 140-150 um; spores colourless, transversely 
septate, 6-7 x 24-33 um. I+. 


CHEMISTRY. Hypoprotocetraric acid and a lower unidentified spot (?4-O-demethylnotatic 
acid). 


HABITAT. Mid bole of trees in rain forest at low elevations (150 m). 
DISTRIBUTION. Sri Lanka. 
ADDITIONAL SPECIMENS. Hale collections: 9 (46 166). 


OBSERVATIONS. Ocellularia kanneliyensis can be recognized by the granular thallus and 
large, emergent apothecia which lack a columella. Another species in Sri Lanka with hypo- 
protocetraric acid, O. neopertusariiformis, has larger apothecia and large spores (80-155 wm 
long). On the world level, O. domingensis (Nyl.) Mull Arg., known from the New World 
(Hale, 1978a : 20), differs in having a smooth thallus, almost urceolate apothecia, and very 
large spores (150-210 wm long). This same evolutionary series includes O. eumorpha (see 
above). 


14. Ocellularia keralensis Patwardhan & C. Kulkarni ex Hale sp. nov. 
(Figs 17e, f) 


Thallus corticola, epiphloeodes, stramineo-albidus, nitidus, verruculosus, aetate rimosus, 8-16 cm 
latus; apothecia dispersa, semi-emergentia, 0°6-1:0 mm diametro, excipulo fuligineo, columella 
nulla; ostiolum rotundatum, c. 0-1 mm diametro, depressum, anguste nigro-cinctum; hymenium 
200-300 um altum; sporae incolores, murales, 25-45 x 100-210 wm, dense loculatae, | : nae, 1+ 
coeruleae. 

Typus: India, Kerala, Munnar-Kodai road below Gellapatti, 1 Jan. 1976, M. B. Nagarkar & K. D. 
Gole 76°542 (AMH—holotypus; US—isotypus). 


Thallus pale tannish gray, 8-16 cm broad, shiny, verruculose, fissured with age; cortex 
irregularly developed and pored, 10-15 um; algal layer continuous, 15 wm; medulla to 
50 um but mostly hypophloeodal; apothecia dispersed, becoming emergent or remaining 
barely semi-emergent, 0°6—1:0 mm diam, the exciple carbonized apically; columella lacking; 


pore round, c. 0°! mm diam, becoming narrowly black rimmed and depressed; hymenium - 


200-300 um; spores colourless, muriform, 25-45 x 100-210 wm, with numerous locules, 
1/ascus, I+. 


CHEMISTRY. No substances present. 


HABITAT. Lower trunks into the canopy of trees in rain forest at low to mid elevations 
(150-1100 m) 


| 


THELOTREMATACEAE IN SRI LANKA 309 
DISTRIBUTION. India and Sri Lanka. 


ADDITIONAL SPECIMENS. Hale collections: 1 (50 215), 3a (46 292, 46 319, 46 326, 46 355), 4 
(47 075, 47 080, 47 087, 47 114), 5 (46 224, 46 231a, 46 225, 46 260, 46 264, 46 269, 46 274, 
46 281), 8 (47 041). 


OBSERVATIONS. While first recognized from India, where it occurs in the Western Ghats, O. 
keralensis is more common in the rain forests of Sri Lanka. It is an inconspicuous species 
with small-pored, sometimes barely raised apothecia. I know of no similar species in the 
genus. 


15. Ocellularia lankaensis Hale sp. nov. 
(Fig. 17g) 


Thallus corticola, epiphloeodes, viridi-glaucescens, 8-16 cm latus; apothecia emergentia, 1-1-3 mm 
diametro, excipulo fuligineo, columella centrali 90-320 wm diametro; ostiolum rotundatum, 
0:2-0:3 mm diametro, pallide croceum; hymenium 90-120 “wm altum; sporae incolores, transversim 
septatae, octonae, 4-7 x 10-22 wm, 5-7 loculatae, I+ coerulaea. 

Typus: Sri Lanka, Western Province, Kalutara District, near Hedigalla, Morapitiya, elev. 150 m, 
11 Feb. 1976, M. E. Hale 47 184 (US—holotypus; BM—isotypus). 


Thallus pale greenish mineral gray, 8-16 cm broad, shiny, smooth to verruculose, fissured 
with age; cortex dense, 10-15 um, with aculeate hyphae, splitting into several layers 
internally; algal layer continuous, 10-15 ~m; medulla 30-60 “m with abundant deposits of 
psoromic acid and oxalate crystals; apothecia numerous, emergent, |—-1°3 mm diam, the 
exciple carbonized; columella 90-320 um diam, the tip and pore area usually distinctly 
orange pruinose, rarely white; pore round, 0:2-0°3 mm diam; hymenium 90-120 um; spores 
colourless, transversely septate, 4-7 x 10-22 um, 5-7 loculate, I+. 


CHEMISTRY. Psoromic acid with or without norpsoromic acid. 
Hasitat. Base, lower trunk, and saplings in rain forest at low to mid elevations (150-850 m). 
DISTRIBUTION. Sri Lanka. 


ADDITIONAL SPECIMENS. Hale collections: 2 (51 106,51 110,51 145,51 153,51 154, 51 180, 
51 184, 51 234), 3a (46 320, 46 334, 46 340, 46 357, 46 366), 3b (46 385, 46 389), 4 (47 094, 
47 095, 47 151, 50 407, 50 412, 50 490, 50 506), 6 (47 181), 7a (SO 163, 50 165, 50 180), 7b 
(51 008, 51 065, 51 249), 8 (47 032), 9 (46 188). Thwaites collections: C. L. 124 (BM, H-Nyl. 
s.n.,S, UPS) (as ‘Let. 81, T. olivaceum’) C. L. 262, 264, 265 (PDA). 


OBSERVATIONS. This species has a conspicuous pale orange-pruinose pore area, although the 
colouration is lacking in a few specimens. A P+ yellow colour test will separate it from O. 
croceopora (‘chonestoma’ unknown present), a much rarer species with similar orange 
pruina. It has larger apothecia than other columellate psoromic acid-containing species such 
as O. antillensis Hale, O. comparabilis (Krempelh.) Mill. Arg., or O. discoidea (Ach.) Mill. 
Arg., none of which occur in Sri Lanka. I have not yet seen any material of O. lankaensis 
outside of Sri Lanka, where it is one of the most common species in the family. 

Both Nylander and Leighton erred in identifying Thwaites C. L. 124 as ‘7. olivaceum 
Mont.’ An isotype of Montagne’s species in Geneve is identical with Myriotrema immersum 
(Eschw.) Hale, which contains protocetraric acid. 


16. Ocellularia leucomelaena (Nyl.) Hale in Mycotaxon 11 : 137 (1980). 
(Fig. 17h) 


Thelotrema leucomelaenum Nyl. in Acta Soc. Scient. fenn. 7: 452 (1863). Type: Colombia, Lindig 
2777 (H-Nyl. 22576—lectotype; BM, FH-Tuck., G—isolectotypes). 

IcONES. Redinger, 1936:92 fig. 55.—Hale, 1974a:37 fig. 17a (lectotype) and fig. 17d.—Hale, 
1978a : 40 fig. 121. 


For additional synonymy see Hale, 1974a: 36. 


310 M. E. HALE 


Thallus whitish mineral gray, c. 6 cm broad, dull, continuous; cortex weakly developed or 
lacking; algal layer 10-15 wm, discontinuous among superficial medullary hyphae; medulla 
mostly hypophloeodal; apothecia immersed, 0°6-0°8 mm diam, exciple carbonized; colu- 
mella very wide, to 360 wm; pore round to irregular, 0°2-0°3 mm, black rimmed, black tip of 
columella clearly visible; hymenium 12 “4m; spores colourless, muriform, 9 x 21-24 wm, 
1-3 x 5-7 loculate, I—. 


CHEMISTRY. No substances present. 

Hasirtat. Branches of trees at higher elevations (2100 m). 

DISTRIBUTION. West Indies, Central and South America, Ivory Coast, Sri Lanka, Hawaii. 
ADDITIONAL SPECIMENS. Hale collections: 10 (50 294). 


OBSERVATIONS. Nylander reported the characteristic I— spore reaction in his original 
description and Redinger (1936 : 91) emphasized the broad columella. Although widespread 
in the New World, where I believed it to be endemic, we now have collections not only from 
Sri Lanka but also from the Ivory Coast (Santesson 10 658e in S). 


17. Ocellularia lirelliformis (Tuck.) Hale in Mycotaxon 11 : 137 (1980). 
(Fig. 171) 


Thelotrema lirelliforme Tuck. in Proc. Am. Acad. Art Sci. 6: 270 (1864). Type: Cuba, Wright 150 
(FH-Tuck.—lectotype; BM, H-Nyl. 22689, G, L, M, PC, UPS, US—isolectotypes). 

Rhabdodiscus lirelliformis (Tuck.) Vainio in Suomal. Tiedeakat. Toim. A 15(6) : 184 (1921). 

Phaeotrema lirelliforme (Tuck.) Zahlbr., Cat. Lich. Univers. 2 : 608 (1923). 


Thallus whitish mineral gray, 4-8 cm broad, dull, continuous to somewhat fissured; cortex 
not fully developed, 10-20 um; algal layer 10-15 wm, interrupted by crystals; medulla 
mostly hypophloeodal; apothecia immersed, round to elongate, 0°8-1:3 mm long, the 
thalline rim suberect, rarely erect and slightly recurved, exciple carbonized; columella 
broad, to 0°6 mm broad, inner wall and disc black or becoming white pruinose; hymenium 
120-150 um; spores brown, transversely septate, 6-9 x 18-21 um, 4 loculate, I—. 


CHEMISTRY. No substances present. 

HaBITAT. Canopy of trees 1n rain forest at lower elevations (300 m). 
DISTRIBUTION. West Indies, Central America, Sri Lanka. 
ADDITIONAL SPECIMENS. Hale collections: 7b (51 035, 51 092). 


OBSERVATIONS. Tuckerman compared this species with Thelotrema auberianum Mont. 
(= Ocellularia auberiana (Mont.) Hale) (see Hale, 1978a: 35), which has psoromic acid 
and a reticulate columella. 


18. Ocellularia marivelensis (Vainio) Hale in Mycotaxon 11 : 137 (1980). 
(Fig. 18a) 


Thelotrema marivelense Vainio in Suomal. Tiedeakat. Toim A, 15(6) : 176 (1921). Type: Philippines, 
Prov. Bataan, Mt Mariveles, Feb. 1905, Whitford 1088 (TUR-Vain. 26 772—lectotype; F, FH, 
US—isolectotypes). 

Leptotrema marivelense (Vainio) Zahlbr., Cat. Lich. Univers. 2 : 637 (1923). 


Thallus ashy green, 6-12 cm broad, appearing thick, shiny, continuous; cortex dense but 
thin, 5-10 wm, with aculeate hyphae, some splitting and exfoliation; algal layer 10-15 um; 
medulla 30-50 wm with numerous crystals; apothecia semi-emergent, 0°8—1-2 mm diam, the 
exciple carbonized; columella variable, 120-310 wm diam, the white pruinose tip visible in 
the pore; pore rather open, 0°2-0'4 mm diam; hymenium 110-130 um; spores brown, 
muriform, 8-12 x 12-20 wm, 1-2 x 4—7 loculate. 


THELOTREMATACEAE IN SRI LANKA 311 
CHEMISTRY. Psoromic acid. 
HapitTAT. Canopy of trees in rain forest at low elevations (150-350 m). 
DISTRIBUTION. India, Sri Lanka, Philippines, Fiji, Solomon Islands. 


ADDITIONAL SPECIMENS. Hale collections: 4 (47 084, 47 123, 47 154, 50 148), 9 (46 130, 
46 164). 


OBSERVATIONS. This Asian species is obviously related to O. crassa and O. fissa (see above) 
but the columella lacks reticulations. 


19. Ocellularia massalongoi [‘massalongi’] (Mont.) Hale in Mycotaxon 11 : 137 (1980). 
(Fig. 18b). 


Ascidium massalongoi [‘Massalongi’] Mont. in Annis Sci. nat. (Bot.) IV, 14: 174 (1860). Type: ‘Ind. 
Or.’ s.c. no. 94 (PC—lectotype). 

Macropyrenium pertusarioides Hampe in Massal., Esame Comp.: 330 (1860). Type: Ceylon, s.c., s.n. 
(H-Nyl. 2244la—lectotype?). 

Ascidium monobactrium Nyl. in Annls Sci. nat. (Bot.) IV, 15:53 (1861). Type: Sabah, Labuan, 
Tonjong Kubong, Motley s.n. (H-Nyl. 22428b—lectotype; BM, PC—isolectotypes). 

Ascidium pachystomum Leighton in Trans. Linn. Soc. Lond. 27: 171 (1870). Type: Sri Lanka, South 
of Island, Thwaites C. L. 171 (BM—lectotype; G, NY, PC, PDA, S, UPS—isolectotypes). 

Thelotrema pachystomum (Leighton) Mill. Arg. in Flora, Jena 74 : 112 (1891). 

Thelotrema monobactrium var. subgranulatum Nyl., Sert. Lich. Trop. Labuan Sing. 6 : (1891). Type: 
Sabah, Labuan, Tonjong Kubong, Motley s.n. (H-Nyl. 22428—lectotype; BM (as no. 2)— 
isolectotype). 

Thelotrema monobactrium (Nyl.) Vainio in Suomal. Tiedeakat. Toim. A, 15(6): 177 (1921). 

Thelotrema marginans Vainio in Suomal. Tiedeakat. Toim. A, 15(6): 179 (1921). Type: Philippines, 
Biliran, June 1914, R. MacGregor Bur Sci. 18422 (TUR-Vain. 26 853—lectotype). 

Thelotrema monobactrium var. endoleucum Vainio in Suomal. Tiedeakat. Toim. A, 15(6): 178 
(1921). Type: Philippines, Prov. Sorsogon, Irosin, Nov. 1915, E/mer 15 050 (TUR—lectotype; BM, 
FH, G, L, LD, US, W—isolectotypes). 

Thelotrema monobactrium var. endorhoda Vainio in Suomal. Tiedeakat. Toim. A, 15(6): 178 
(1921). Type: Philippines, Prov. Sorsogon, Irosin, Nov. 1915, E/mer 15 078 (TUR—Vain. 26 826— 
lectotype; FH, L, US—isolectotypes). 

Thelotrema monobactrium var. geminipara Vainio in Suomal. Tiedeakat. Toim. A, 15(6) : 178 (1921). 
Type: Philippines, Prov. Sorsogon, Irosin, April 1916, E/mer 14 813 (FH, US—isotypes). 

Thelotrema leightonii Zahlbr., Cat. Lich. Univers. 2 : 620 (1923). Based on T. pachystomum (Leighton) 
Mull. Arg., non 7. pachystomum Ny\. 

Thelotrema massalongoi {‘Massalongii’| (Mont.) Zahlbr., Cat. Lich. Univers. 2 : 624 (1923). 


Thallus dull greenish mineral gray, 8-14 cm broad, shiny, smooth and continuous; cortex 
dense, very thick, 30-40 wm, with aculeate hyphae, splitting internally into layers; algal 
layer continuous, 15-17 wm; medulla to 100 um with crystals; apothecia strongly emergent, 
solitary or less commonly aggregated in twos, |-1:3 mm diam, exciple carbonized; columella 
c. 300 um diam; pore round, 0:08-0:1 mm diam, depressed but surrounded by a strongly 
raised rim; hymenium 300-400 um; spores colourless, muriform 25-30 x 150-360 um, with 
numerous locules, 1 /ascus, I+. 


CHEMISTRY. Salazinic acid. 


HasitaT. Saplings and lower trunk to mid bole of trees in rain forest at lower elevations 
(150-350 m). 


DISTRIBUTION. Sri Lanka, Malaya, Sabah, Philippines, Japan, Solomon Islands. 


ADDITIONAL SPECIMENS. Hale collections: 3a (46 285, 46 361), 4 (SO 376, 50 484), 7a (50 144, 
50 160, 50 167, 50 173). Thwaites collections: C. L. 170 (BM, PDA) (as ‘Lgt. 99. A. pachy- 
stomum’). 


OBSERVATIONS. Ocellularia massalongoi is easily recognized by the large apothecia with a 


312 M. E. HALE 


Fig. 18 Species of Ocellularia. (a) O. marivelensis (Hale 46 130). (b) O. massalongoi (Hale 50 484). 
(c) O. meiosperma (Almquist s.n. in S; isolectotype of Thelotrema subinalbescens). (d) O. 
meiospermoides (Hale 46 268). (e) O. melanotremata (Hale 46 241). (f) O. monosporoides (Hale 
50 276). (g) O. neocavata (Hale 50 255). (h) O. neopertusaritiformis (Hale 47 145). () O. 
nylanderiana (Hale 47 132). See Fig. 71 for scale. 


THELOTREMATACEAE IN SRI LANKA 313 


raised rim around the pore, a trait mentioned by all earlier workers, and the chemistry. 
Salazinic acid is extremely rare in the family, known as far as I am aware in just Ocellularia 
interponenda (Nyl.) Hale, O. massalongoi and Thelotrema hians Stirton. 


20. Ocellularia meiosperma (Nyl.) Hale in Mycotaxon 11 : 137 (1980). 
(Fg. 17c) 


Thelotrema meiospermum Nyl. in Annls Sci. nat. (Bot.) 1V, 19 : 333 (1863). Type: Cuba, Wright 136 
(FH-Tuck.—lectotype; BM, G, H-Nyl. 22858, L, PC, UPS—isolectotypes). 

Phaeotrema meiospermum (Nyl.) Mill. Arg. in Flora, Jena 69 : 311 (1886). 

Thelotrema subinalbescens Nyl. in Acta Soc. Scient. fenn. 26 : 18 (1900). Type: Sri Lanka, Pointe de 
Galle, A/mquist s.n. (S—lectotype). 

Verrucaria discissa Nyl. in Acta Soc. Scient. fenn. 26 : 25 (1900). Type: Sri Lanka, Pointe de Galle, 
Almquist s.n. (H-Nyl. 1361—lectotype). 

Pyrenula discissa (Nyl.) Zahlbr., Cat. Lich. Univers. 1 : 428 (1921). 

Phaeotrema subinalbescens (Nyl.) Zahlbr., Cat. Lich. Univers. 2 : 610 (1923). 

For additional synonymy see Hale, 1978a : 36. 

Icongs. Hale, 1978a: 35 fig. 9e. 


Thallus whitish mineral gray, 1-2 cm broad, dull, continuous; cortex not clearly developed, 
a thin polysaccharide layer covering the surface; algal layer scattered among medullary 
hyphae; medulla 25-30 um, with crystals; apothecia immersed, round to irregular, 
0:2-0'4 mm diam, the exciple carbonized; columella broad, to 0°3 mm diam, becoming 
somewhat reticulated at maturity; hymenium 90-100 um; spores brown, transversely 
septate, 10 x 16-20 um, 4-S loculate, I—. 


CHEMISTRY. No substances present. 
HaBiTAT. Secondary forest at low elevations (100 m). 


DISTRIBUTION. United States, West Indies, Central America, Sri Lanka, Philippines, Java, 
Sarawak. 


OBSERVATIONS. I did not collect this rather inconspicuous lichen in Sri Lanka. It is 
apparently confined to secondary lowland forests. Nylander described Thelotrema subinal- 
bescens at the end of his career, saying ‘affine Th. phaeospermum Ny|l. in Wright. Cub. n: o 
45’, apparently an otherwise unpublished name. I have had no success in tracing the identity 
of the Cuban specimen. Dr P. G. Patwardhan (Poona) was the first to recognize Verrucaria 
discissa as a Phaeotrema, and he called my attention to it. 


21. Ocellularia meiospermoides Hale sp. nov. 
(Fig. 18d) 


Thallus corticola, hypophloeodes, opacus, 8—10 cm latus; apothecia emergentia, solitaria, 0°6-1‘0 mm 
diametro, excipulo fuligineo, columella 60-180 ~m diametro; ostiolum rotundatum, 0°1-0°3 mm 
diametro, albo-cinctum; hymenium 110-130 4m altum; sporae fuscae, transversim septatae, 
octonae, 6-7 x 24-33 yum, 6-7 loculatae, I—. 

Typus: Sri Lanka, Southern Province, Galle District, Kanneliya Forest Reserve, elev. 150 m, 16 
Feb. 1976, M. E. Hale 46 268 (US—holotypus; BM, PDA—isotypi). 


Thallus tannish mineral gray, 8-10 cm broad, dull, thin, continuous; cortex lacking; algae 
scattered among superficial medullary hyphae; medulla mostly hypophloeodal; apothecia 
emergent, solitary, 0-6-1:0 mm diam, the exciple carbonized; columella 60-180 um; pore 
round, 0:1-0°3 mm diam, white rimmed; hymenium 110-130 um; spores brown, trans- 
versely septate, 6-7 x 24-33 um, 6-7 loculate, I—. 


CHEMISTRY. No substances present. 
HasitTAtT. Lower trunks of trees in rain forest at lower elevations (150-850 m). 


DISTRIBUTION. Sri Lanka. 


314 M. E. HALE 


ADDITIONAL SPECIMENS. Hale collections: 2 (50 161). 


OBSERVATIONS. I am describing this species on the basis of the distinctly emergent apothecia 
and the columella higher than wide. In other respects it is close to O. meiosperma (above). 


22. Ocellularia melanotremata Hale sp. nov. 
(Fig. 18e) 


Thallus corticola, hypophloeodes, opacus, albo-cinereus, 4-12 cm latus; apothecia immersa vel vix 
semi-emergentia, 0°2-0'3 mm diametro, excipulo fuligineo, columella 45-60 4m; ostiolum 
rotundatum, 0:05-0:10 diametro, nigro-cinctum; hymenium 80-120 um altum; sporae incolores, 
transversim septatae, 6-10 x 15-24 wm, 1-2 x 5-7 loculatae, 4-8 : nae, I—. 

Typus: Sri Lanka, Southern Province, Matara District, mossy forest above Enselwatta, elev. 
1100 m, 14 Feb. 1976, M. E. Hale 46 241 (US—holotypus; BM—isotypus). 


Thallus whitish gray, 4-12 cm broad, very thin, dull, continuous; cortex lacking; algae 
scattered among superficial medullary hyphae; medulla mostly hypophloeodal; apothecia 
more or less immersed to barely emergent, 0°2-0°3 mm diam, the exciple carbonized; 
columella 45-60 wm diam; pore round, 0°05-0'1 mm diam, becoming black rimmed; 
hymenium 8-120 wm; spores colourless, transversely septate, 6-10 x 15-24 wm, 1-3 x 5-7 
loculate, 4—-8/ascus, I—. 


CHEMISTRY. Stictic acid. 


HaBITAT. Saplings, base or lower trunks of trees in low to mid elevation rain forest 
(150-1100 m). 


DISTRIBUTION. Sri Lanka. 


ADDITIONAL SPECIMENS. Hale collections: 1 (50 241, 50 242), 2 (51 152), 4 (47 — 50 413, 
50 425, 50 472), 7a (50 169, 50 172, 50 174), 7b (51 012). 


OBSERVATIONS. This ecorticate species is very close to O. erumpens (Magnusson) Hale (see 
Hale, 1978a: 43), which has larger spores (36-50 um long) and apothecia (0°5-0°7 mm 
diam). It is rather common in the rain forests of Sri Lanka and will undoubtedly be found in 
the Indo-Malaysian area by future collectors. 


23. Ocellularia monosporoides (Nyl.) Hale in Mycotaxon 11 : 137 (1980).° 
(Fig. 18f) 


Thelotrema monosporoides Nyl., Lich. New Zealand: 76 (1888). Type: New Zealand, Knight s.n. 
(H-Nyl.—lectotype; G—isolectotype). 

Leptotrema monosporoides (Nyl.) Miill Arg. in Bull. Herb. Boissier 2 (append. 1) : 75 (1894). 

Thallus pale straw coloured, 2-6 cm broad, appearing thick, dull, smooth to verruculose, 
fissured; cortex dense, 15-20 wm; algal layer continuous, 10-15 wm; medulla 10-15 wm, 
with crystals, mostly hypophloeodal; apothecia semi-emergent, 1-1-3 mm diam, the exciple 
carbonized apically; columella lacking; pore variable, round, 0-1-0°3 mm diam, more or less 
depressed; hymenium 280-300 um; spores nearly colourless to pale brown or darkening at 
maturity, muriform, 30-35 x 120-210 wm, with numerous locules, 1/ascus, I—. 


CHEMISTRY. No substances present. 
HaBITAT. Trees in open forest at higher elevations (2100-2200 m). 
DISTRIBUTION. Sri Lanka, New Zealand. 


ADDITIONAL SPECIMENS. Hale collections: 10 (50 262, 50 265, 50 266, 50 276, 50 278, 50 
285, 50 289, 50 291), 12 (50 346, 50 360). Thwaites collections: C. L. 141 (BM, PDA, Ss; 
UPS) (as ‘Lgt. 95. Ascidium depressum’). Other collections: Horton Plains, Moberg 2619 
(UPS), Santesson 25 862, 25 941, 26 028, 26 031 (S). 


THELOTREMATACEAE IN SRI LANKA 315 


OBSERVATIONS. Nylander separated this species from Thelotrema monosporum by the larger 
spores (140-200 um long) in his original description. A more basic difference, of course, is 
that O. monosporoides has a carbonized exciple and lacks periphysoids. The spores may be 
colourless well into the mature stages, and only slowly and imperfectly turn brown in Sri 
Lanka. Leighton misidentified the species as ‘Ascidium depressum’ and failed to send a 
duplicate to Nylander for verification; Thelotrema depressum Mont. (see Hale, 
1974a : 33) isa heavily carbonized, columellate species from the New World. 


24. Ocellularia neocavata Hale sp. nov. 
(Fig. 18g) 


Thallus corticola, epiphloeodes, rugulosus, viridi-cinerascens, 8—12 cm latus; apothecia immersa, vel 
vix semi-emergentia, 0°3-0°4 mm diametro, excipulo pallide fuligineo, columella 60 um diametro; 
ostiolum rotundatum, 0°! mm diametro; hymenium 90 um altum; sporae incolores, transversim 
septatae, octonae, 5 x 15-25 wm, 7-8 loculatae, I+ coeruleae. 

Typus: Sri Lanka, Sabaragamuwa Province, Kegalla District, Maliboda to Theberton, elev. 800 m, 
16 Mar. 1978, M. E. Hale 50 255 (US—holotypus; BM—isotypus). 


Thallus pale greenish mineral gray, 8-12 cm broad, shiny, verruculose and rugulose, 
fissured; cortex dense, 15 um; algal layer 15 wm; medulla to 150 wm under verrucae, with 
large crystals; apothecia numerous, immersed, flush to somewhat raised, 0°3-0°4 mm diam, 
the thalline rim jagged and white pruinose apically, exciple mostly fused, the tips filling the 
pore; columella 60 wm diam, weakly carbonized; pore round to irregular, partially closed, 
0:1 mm diam; hymenium 90 ym; spores colourless, transversely septate, 5 x 15-25 wm, 7-8 
loculate, I+. 


HaBiTAT. Saplings along trail in rain forest at lower to mid elevations (300-800 m). 
DISTRIBUTION. Sri Lanka. 
ADDITIONAL SPECIMENS. Hale collections: 7b (51 240). 


OBSERVATIONS. This species differs from O. cavata (Ach.) Mill. Arg. in having a semi-erect, 
white pruinose thalline rim with the apothecia barely raised. The walls and the columella are 
weakly carbonized. By comparison, O. cavata has emergent apothecia with a small, discrete 
pore and heavily carbonized exciple and columella. The chemistries are identical, a series of 
H,SO,+ gray spots between fumarprotocetraric and protocetraric acids. In Sri Lanka O. 
punctulata also has this chemistry and at the world level O. cavata, O. crocea (Krempelh.) v. 
Overeem & D. v. Overeem, O. decolorata Hale, O. dodecamera (Nyl.) Vainio, O. exigua 
Mill. Arg., O. polydisca Redinger, O. rhabdospora (Nyl.) Redinger, O. subemersa Miill 
Arg., O. xanthostromiza (Nyl.) Zahlbr., and one species with muriform spores, O. 
neodominicana Hale (see Hale, 1974a : 33, under Thelotrema dominicanum). No species of 
Myriotrema or Thelotrema are known to contain this substance. 


25. Ocellularia neopertusariiformis Hale sp. nov. 
(Fig. 18h) 


Thallus corticola, epiphloeodes, albo-viridis vel cinereus, 8-12 cm latus; apothecia valde emergentia, 
basin constricta, 1°5—1-9 mm diametro, excipulo fuligineo, columella nulla; ostiolum rotundatum, 
0:2-0:4 mm diametro, nigro-cinctum; hymenium 210 um altum; sporae incolores, transversim 
septatae, 12-22 x 80-155 wm, 24-37 loculatae, 2-4 : nae, I+ coeruleae. 

Typus: Sri Lanka, Sabaragamuwa Province, Ratnapura District, Sinharaja Forest Reserve, 
Weddagala, 13 Feb. 1976, M. E. Hale 47 145 (US—holotypus; AMH, BM, PDA, isotypi). 


Thallus greenish to whitish mineral gray, 8-12 cm broad, smooth to rugulose, continuous; 
cortex loosely organized and irregularly pored, c. 5 wm thick; algal layer continuous, 10 um; 
medulla 10 wm; mostly hypophloeodal; apothecia conspicuous, strongly emergent with a 
more or less constricted base, 1:5—1:9 mm diam, the exciple carbonized; columella lacking; 


pore round to irregular, depressed at maturity, 0:2-0°4 mm diam, black rimmed; hymenium 


316 M. E. HALE 


210 um; spores colourless, transversely septate, 12-22:x:80=155 um, 24-27 loculate, 
2-4/ascus, I+. 


CHEMISTRY. Hypoprotocetraric acid. 


HABITAT. Lower trunk to mid bole of trees in rain forest at lower to mid elevations 
(300-800 m). 


DISTRIBUTION. Sri Lanka. 
ADDITIONAL SPECIMENS. Hale collections: 2 (51 230), 4 (47 120, 47 146, 50 395). 


OBSERVATIONS. I had at first identified this species with O. pertusariiformis but the 
chromatographic profiles differ, although apparently very closely related compounds are 
involved. Aside from the chemical difference, O. neopertusariiformis has a larger, depressed 
pore and a thinner, pored cortex. It also occurs at lower elevations, whereas rare O. pertu- 
sariiformis seems to be restricted to high elevations. 


26. Ocellularia nylanderiana Hale in Phytologia 26 : 414 (1973). 
(Fig. 181) 


Ascidium majorinum var. longius Nyl., Sert. Lich. Trop. Labuan Singapore: 20 (1891). Type: 
Singapore, 1879, Almquist s.n. (H-Nyl. 22 386—lectotype; S—isolectotype). 
Icongs. Hale, 1973 : 41 fig. 3. 


Thallus whitish to ashy gray, 3-12 cm broad, shiny, becoming rugulose-verruculose; cortex 
loosely organized with irregular pores, 8-12 um; algal layer continuous, 15 wm; medulla 
10-15 um, with crystals, mostly hypophloeodal; apothecia not abundant, strongly emergent, 
1-1-5 mm diam, the amphithecium rugose, easily abraded to expose the underlying thick 
medulla, exciple heavily carbonized; columella 200-290 um diam; pore irregular, 
0:2-0'-4 mm diam, sometimes faintly white rimmed; hymenium 100-240 wm; spores 
colourless, transversely septate, 10-20 x 90-180 wm, about 18 loculate 2/ascus, I+. 


CHEMISTRY. ‘Chonestoma’ unknowns (2 P— spots). 
HaBITAT. Lower trunk into the canopy in rain forest at low elevations (100-300 m). 
DISTRIBUTION. Sri Lanka, Peninsular Malaysia, Philippines, Solomon Islands. 


ADDITIONAL SPECIMENS. Hale collections: 3a (46 317), 4 (47 096, 47 132, 47, 139, 50 474), 6 
(47 188), 7a (50 096, 50 153), 7b (51 088, 51 095, 51 246, 51 256), 8 (47 021, 47 038, 47 044, 
47 045), 9 (46 203). 


OBSERVATIONS. I have placed under this name all columellate, ‘chonestoma’ unknown- 
containing species with large spores. The far commmoner O. chonestoma has much smaller 
spores, 18-30 wm long, but is otherwise externally very similar. 


27. Ocellularia orthomastia (Krempelh.) Zahlbr., Cat. Lich. Univers. 2 : 597 (1923). 
(Fig. 19a) 


Ascidium orthomastium Krempelh. in Nuovo G. bot. ital. 7 : 60 (1875). Type: Singapore, Beccari 247 
(M—lectotype; W—isolectotype). 

Thallus greenish to whitish gray, 8-12 cm broad, shiny, smooth to verruculose, fissured; 
cortex rather thin and irregularly pored, 5 wm; algal layer 15-20 wm; medulla to 10 wm, with 
some crystals, mostly hypophloeodal; apothecia rather sparsely developed, semi-emergent to 
emergent, 0°8-1:1 mm diam, exciple carbonized; columella c. 190 u~m diam; pore round to 
irregular, 0°3-0-4 mm diam, the tip of the columella visible within; hymenium 160 um; 
spores colourless, transversely septate, 10 x 40-65 um, 12-15 loculate, I+. 


CHEMISTRY. No substances present. 


HABITAT. Lower trunks to canopy in rain forest at low elevations (150 m). 


THELOTREMATACEAE IN SRI LANKA 317 


DISTRIBUTION. Sri Lanka, Malaya, Philippines, Sarawak, Sabah. 
ADDITIONAL SPECIMENS. Hale collections: 9 (46 142, 46 156, 46 192). 


OBSERVATIONS. This is the first report of the species since its original publication in 1875. 
Krempelhuber could not find spores but cited measurements (7 x 52 4m) made by the 
sender, Dr Koch Onosbalci. Most apothecia in the type collections are indeed moribund, but 
I found a fertile one, confirming the measurements by Koch. The spores are I+ blue and 
12-15 loculate. 


28. Ocellularia papillata (Leighton) Zahlbr., Cat. Lich. Univers. 2 : 597 (1923). 
(Fig. 19b) 


Thelotrema papillatum Leighton in Trans. Linn. Soc. Lond. 27 : 169 (1870). Type: Sri Lanka, Central 
Province, Thwaites C. L. 129 (BM—lectotype; H-Nyl. 3854, G, PC, PDA, S, UPS, W— 
isolectotypes). 

IcONES. Hale, 1974a : 23 fig. 12d (lectotype) and fig. 12e—Hale, 1978a : 27 fig. 7b. 


Thallus whitish to light greenish gray, 3-12 cm broad, shiny, continuous, smooth to verrucu- 
lose with age; cortex dense, 20-25 wm, with poorly developed aculeate hyphae, splitting 
somewhat internally; algal layer continuous, 10 “4m; medulla to 10 um, mostly hypo- 
phloeodal; apothecia common, immersed, sometimes slightly raised, 0°-4—0:7 mm diam, the 
exciple becoming carbonized; columella variably developed, 20-210 um, rarely lacking; 
pore round, 0°1-0:15 mm diam, becoming white rimmed; hymenium 80-160 um; spores 
colourless, transversely septate, 4-6 x 12-28 wm, 5-6 loculate, I+. 


CHEMISTRY. No subsances present. 
HaBiTAT. Saplings, base to mid bole of trunks in rain forest at lower elevations (150-850 m). 


DISTRIBUTION. West Indies, Central and South America, India, Sri Lanka, Philippines, New 
Caledonia. 


ADDITIONAL SPECIMENS. Hale collections: | (50 224), 2 (51 107), 3a (46 288), 3b (46 390), 4 
(47 101, 47 113, 47 163, 50 380, 50 390, 50 396, 50 419, 50 424, 50 427, 50 431, 50 435, 50 
436, 50 437, 50 438, 50 445, 50 450, 50 451, 50 455, 50 460, 50 460A, 50 462, 50 471, 50 
473), 7a (50 140, 50 171), 7b (50 309, 51 030, 51 062), 8 (47 004, 47 050), 9 (46 216). 
Thwaites collections: C. L. s.n. (PDA). 


OBSERVATIONS. As in Panama (Hale 1978a : 26), this is one of the commonest species at base 
level in lower elevation rain forest. The columella varies considerably, from weakly if at all 
developed, to distinct. In general the Sri Lankan specimens only rarely have raised 
apothecia. 


29. Ocellularia perforata (Leighton) Mill. Arg. in Hedwigia 31 : 284 (1892). 
(Fig. 19c) 


Thelotrema perforatum Leighton in Trans. Linn. Soc. Lond. 25 : 447 (1886). Type: Brazil, Casiquiari, 
Spruce 254 (BM—lectotype; H-Nyl. 22557—isolectotype). 

For additional synonymy see Hale, 1974a: 25. 

ICONES. Redinger, 1936 : 50 fig. 26.—Hale, 1974a : 23 fig. 12f (lectotype)—Hale 1978a : 37 fig. 7c. 


Thallus pale yellowish to greenish gray, 8-10 cm broad, shiny, smooth to verruculose; cortex 
dense but rather thin, 5-7 wm, with weakly developed aculeate hyphae; algal layer 
continuous, 15 wm; medulla 10 wm or less, mostly hypophloeodal; apothecia numerous, 
immersed to barely semi-emergent, 0°2-0°4 mm diam, the exciple carbonized; columella 
45-100 um diam; pore round, 0:05-0'1 mm diam, often filled with tip of columella; 
hymenium 75-120 um; spores colourless, transversely septate, 4-6x 12-24 um, 5-8 
loculate, I+. 


CHEMISTRY. Protocetraric acid. 


318 M. E. HALE 


“oe é be ‘ / va 4 : } ay J ; y 4, 
Fig. 19 Species of Ocellularia. (a) O. orthomastia (Hale 46 192). (b) O. papillata (Hale 445). (c) O. 
perforata (Hale 51 546). (d) O. pertusariiformis (Hale 50 274). (e) O. pluripora (Hale 51 168). (f) 


O. polillensis (Hale 50 510). (g) O. punctulata (Hale 51 243). (h) O. pyrenuloides (Hale 46 283). 
(i) O. rassagala (Hale 51 216). See Fig. 71 for scale. 


THELOTREMATACEAE IN SRI LANKA 319 
HABITAT. Saplings and base of trees in rain forest at low to mid elevations (150-850 m). 


DISTRIBUTION. West Indies, Central and South America, India, Sri Lanka, Java, Australia, 
Solomon Islands, Hawaii. 


ADDITIONAL SPECIMENS. Hale collections: 1 (50 220), 2 (51 101), 3b (46 387), 4 (47 103, 51 
112,51 146,51 159,51 176,51 182,51 183), 7b(51 014, 51 028, 51 071), 8 (47010). 


OBSERVATIONS. Ocellularia perforata occupies the same habitats as O. papillata and is 
differentiated from it by the chemisty (P+ red) and the generally semi-emergent apothecia. It 
is much less common here than O. papillata; the reverse is true in the American tropics. 
When examining the Spruce type from South America, Redinger (1936 : 51) did not find a 
columella, but I was able to detect one in the BM material, rather weakly developed but 
definitely present. 


30. Ocellularia pertusariiformis (Leighton) Zahlbr., Cat. Lich. Univers. 2 : 598 (1923). 
(Fig. 19d) 


Thelotrema pertusariaeforme Leighton in Trans. Linn. Soc. Lond. 27 : 170 (1870). Type: Sri Lanka, 
Central Province, Thwaites C. L.97 (BM—lectotype; G, PC—isolectotypes). 


Thallus whitish mineral gray, 3-5 cm broad, smooth to verruculose, continuous or fissured 
with age; cortex cellular, 7-10 wm, with some aculeate hyphae and internal splitting; algal 
layer continuous, 15 wm; medulla 10-60 wm with some crystals; apothecia conspicuous, 
emergent and basally constricted, warty, |-1‘4 mm diam, the exciple fused, reddish brown; 
columella lacking; pore area flattened to depressed, pore 0:05-0:08 mm diam, white 
rimmed; hymenium 300-450 um; spores colourless, transversely septate, 28-33 x 105- 
135 wm, 11-12 loculate, 1-2/ascus, I+. 


CHEMISTRY. Unidentified OP-1 and OP-2 (OP-1 near notatic acid and could be the depsi- 
done of evernic acid—C. F. Culberson), both P—. 


HABITAT. Twigs and lower trunk in high elevation rain forest (2100 m). 
DISTRIBUTION. Sri Lanka. 


ADDITIONAL SPECIMENS. Hale collections: 10 (50 274), 12 (50 342). Thwaites collections: C. 
L. 26 (BM, G, H-Nyl. 3922, PC, PDA), C. L. 257 (PDA) (as ‘Lgt. 96, Ascidium pertusariae- 
forme’) (some duplicates in BM and PDA include mixtures of Myriotrema microstomum). 


OBSERVATIONS. This is a rare species occurring only at higher elevations. It is unknown out- 
side of Sri Lanka. The closest relative is O. neopertusariiformis which has different chemistry 
and a more loosely organized cortex. 


31. Ocellularia pluripora Hale sp. nov. 
(Fig. 19e) 


Thallus corticola, epiphloeodes, viridi-cinereus, 5-12 cm latus; apothecia semi-emergentia, solitaria 
vel aggregata, multicavata, 0°6-1(0 mm diametro, excipulo fuligineo, columella 90-130 um 
diametro; ostiolum rotundatum, 0:05-0'1 mm diametro; hymenium 90-120 um altum; sporae 
incolores, transversim septatae, octonae, 5-8 x 14-21 wm, 5-7 loculatae, I+ coeruleae. 

Typus: Sri Lanka, Sabaragamuwa Province, Ratnapura District, Halwathura to Kanda, near 
Rassagala, elev. 900 m, 15 Mar. 1978, M. E. Hale 51 168 (US—holotypus; BM—isotypus). 


Thallus dull greenish gray, 5-12 cm broad, shiny, continuous; cortex dense 10-12 um, with 
aculeate hyphae, splitting internally; algal layer 15 wm, interrupted by crystals; medulla to 
130 wm with remnants of cortical layers and large crystals; apothecia numerous, semi- 
emergent, solitary or fused with 2 or more pores in the same wart, 0°6-1:0 mm diam, the 
exciple weakly carbonized; columella 90-130 um; pore round, 0:05-0:1 mm diam, white 
rimmed; hymenium 90-120 um; spores colourless, transversely septate, 5-8 x 14-21 um, 
5-7 loculate, I+. 


320 M. E. HALE 

CHEMISTRY. Psoromic and norpsoromic acids. 

HasiTAT. Lower trunk into the canopy in rain forest at low to mid elevations (150-1000 m). 
DISTRIBUTION. Sri Lanka. 


ADDITIONAL SPECIMENS. Hale collections: | (50 207), 2 (51 133, 51 204), 3a (46 377), 4 (47 
100, 47 143), 5 (46 225, 46 238), 8 (47 023), 9 (46 919, 46 200). 


OBSERVATIONS. The diagnostic feature of this endemic species is the frequent occurrence of 
grouped or fused apothecia with 1-3 pores. The apothecia are significantly larger and more 
emergent than those of O. terebrata (Ach.) Mill. Arg., the only comparable species on the 
world level. 


32. Ocellularia polillensis (Vainio) Hale comb. nov. 
(Fig. 19f) 


Thelotrema polillense Vainio in Suom. Tiedeakat. Toim. A, 15(6): 180 (1921). Type: Philippines, 
Polillo Island, Aug. 1909, C. Robinson, Bur. Sci. 9098 (TUR-Vain. 26 839—lectotype; 
FH—isolectotype). 

Thallus greenish mineral gray, 6-15 cm broad, shiny, continuous or fissured; cortex dense, 

10-13 wm, with aculeate hyphae, very little splitting internally; algal layer continuous, 

15 um; medulla 10-70 um, with crystals; apothecia semi-emergent to emergent, 0°7-1'1 mm 

diam, exciple carbonized; columella entire but becoming more or less reticulate at maturity, 

320-350 um diam, the pruinose tip often filling the pore; pore gaping, 0°2-0°4 mm diam; 

hymenium 70-120 wm; spores colourless, muriform, 6-10 x 15-24 wm, 1-2 x 4-6 loculate, 

I+. 


CHEMISTRY. Psoromic acid with or without norpsoromic acid. 
HaBIrTAT. Base and lower trunk of trees in rain forest at low to mid elevations (150-850 m). 
DISTRIBUTION. Sri Lanka, New Caledonia, Philippines. 


ADDITIONAL SPECIMENS. Hale collections: 2 (51 184), 3a (46 286, 46 287), 3b (46 383), 4 (47 
137, 50 374, 50 465, 50 467, 50 508), 5 (46 226, 46 252), 6 (47 191), 7b (51 089), 8 (50 508), 9 
(46 154). 


OBSERVATIONS. This common species can be recognized by the gaping pore and reticulate 
columella. Externally it is close to O. crassa, which would be separated by the more erect, 
jagged thalline rim. The psoromic acid-containing species of Ocellularia often present 
taxonomic difficulties since they tend to intergrade. I am recognizing O. polillensis as the 
reticulate-columelilate member of a columellate group that includes O. asiatica (Vainio) 
Hale, O. straminea (Vainio) Hale, and perhaps O. leucina (Mill Arg.) Hale, all of which 
have a simple columella. 


33. Ocellularia punctulata (Leighton) Zahlbr., Cat. Lich. Univers. 2 : 599 (1923). 
(Fig. 19g) 


Ascidium punctulatum Leighton in Trans. Linn. Soc. Lond. 27 : 171 (1870). Type: Sri Lanka, Central 
Province, Thwaites C. L. 127 (BM—lectotype; G, PC, PDA—isolectotypes). 


Thallus light yellowish to greenish gray, 2-8 cm broad, smooth to finely verruculose, fissured 
with age; cortex thin with irregular pores, 5 um, with some aculeate hyphae; algal layer 
10 um; medulla variable 10-30 wm, more or less uniformly orange, with numerous crystals; 
apothecia numerous, emergent, 0°8-1‘(0 mm diam, the exciple carbonized; columella 
70-240 um, protruding into the pore; pore round, somewhat depressed, 0°1-0:2 mm diam; 
hymenium 80-115 um; spores colourless, transversely septate, 5—6 x 15-30 um, 6-10 
loculate, I+. 


CHEMISTRY. ‘Cinchonarum’ unknowns (P+ red) an an unidentified K+ purplish pigment. 


THELOTREMATACEAE IN SRI LANKA Sy) 


HABITAT. Base and lower trunk and saplings in rain forest at low to mid elevations 
(150-850 m). 


DISTRIBUTION. Sri Lanka. 


ADDITIONAL SPECIMENS. Hale collections: | (50 201, 50 212, 50 251), 2 (51 123, 51 191, 51 
197), 3a (46 367, 46 374), 4 (47 070, 47 088, 47 118, 47 122, 50 406, 50 433, 50 440, 50 452, 
50 461, 50 517), 6 (47 170), 7a (50 149, 50 181, 50 185), 7b (50 308, 51 010, 51 041, 51 048, 
51 064, 51 091, 51 243), 8 (47 013, 47 017, 47 019, 47 047), 9 (46 153, 46 178). 


OBSERVATIONS. This is one of the most unusual Sri Lankan endemics, widely distributed in 
the lowland rain forest. The pigmented medulla was not noticed by Leighton. On the world 
level it is a close relative of O. cavata (Ach.) Mill Arg., which has a white or pale yellowish, 
K— medulla and spores generally more than 30 wm long (Hale, 1978a: 16). They have the 
same P+ red chemistry. 

The epithet ‘punctulatum’ has been a source of confusion ever since Leighton published 
the species. Zahlbruckner (1923 : 599) listed it as ‘pustulata’, an obvious slip. Leighton had 
sent it to Nylander for comment in 1868 labelled ‘Ascidium punctulatum’,. Nylander (in litt. 
27 June 1868) identified it as ‘Thelotrema punctulatum Nyl. Lich. Extot. 222’ except that the 
spores were not brown. In the Nylander herbarium at Helsinki I found a specimen labelled 
‘Thelotrema punctulatum Nly., Ins. Cubae, 22446. “Th. 25” ’, which can be identified as 
Myriotrema compunctum. This identification was apparently never published. Nylander 
had, however, already published the epithet without a description in his Conspectus generis 
Thelotrematis (1862 : 96) as ‘27. Th punctulatum Nyl. (Por. compunctum Ach.).—Amer.’ 
The specimen on which this name was based is Nyl.-herb. 22691, which is also Myriotrema 
compunctum Ach., which is, according to Salisbury (1978 : 414), Thelotrema sordidescens 
(Fée) Nyl. At the same time Salisbury lists in synonymy an ‘Ascidium punctulatum (Nyl.) 
Kremph.’, but in his monograph of Ascidium Krempelhuber (1877 : 132) cited ‘Ascidium 
punctulatum Leight.’ and does not refer to the Nylander epithet. In any event the 
combination Ocellularia punctulata appears to be valid. 


34. Ocellularia pyrenuloides Zahlbr. in Magnusson & Zahlbr. in Ark. Bot. 31A(1): 46 
(1944). 
(Fig. 19h) 

Type: Hawaii, Wailuku, Faurie 676 (W—lectotype). 
Icones. Hale, 1974a : 23 fig. 12j (lectotype) and fig. 12k. 
Thallus whitish gray, 8-12 cm broad, shiny, continuous; cortex poorly developed and in 
large part lacking; algal layer scattered among superficial medullary hyphae or hypo- 
phloeodal, to 15 um; medulla mostly hypophloeodal; apothecia immersed in periderm, 
inconspicuous, 0°3 mm diam, the exciple carbonized apically; columella 60-90 um; pore 
round, 0°6-0:9 mm diam, usually darkening; hymenium 40-45 um; spores colourless, trans- 
versely septate, S—6 x 15-18 wm, 4-6 loculate, I+. 


CHEMISTRY. Stictic acid with or without constictic acid. 

HABITAT. Lower trunk in rain forest at low to mid elevations (300-1100 m). 
DISTRIBUTION. West Indies, Venezuela, Hawaii, Sri Lanka. 

ADDITIONAL SPECIMENS. Hale collections: 4 (50 448), 5 (46 283). 


OBSERVATIONS. I had earlier identified this lichen from Dominica in the West Indies (Hale, 
1974a : 25) and it now appears to be pantropical. The blackening of the pore was described 
by Zahlbruckner as ‘circum discum nigro-annulatae’. This is caused by exposure of the 
exciple but can be variable with thalline tissue sometimes masking the rim completely. The 
diagnostic characters would have to be lack of a cortex and presence of a columella and 


By) M. E. HALE 


stictic acid. A very close relative, O. tenuis (below), has somewhat larger, muriform spores 
which are negative with iodine. 


35. Ocellularia rassagala Hale sp. nov. 
(Fig. 191) 


Thallus corticola, epiphloeodes, nitidus, continuus, cinereo-viridis, c. 10 cm latus; apothecia immersa, 
0:3-0:'4 mm diametro, excipulo fuligineo, columella c. 60 um diametro; ostiolum rotundatum, 
0:1 mm diametro, albo-cinctum; hymenium 120-150 wm altum; sporae fuscae, transversim septatae, 
octonae, 6-/ x 14-18 wm, 5-6 loculatae, I—. 

Typus: Sri Lanka, Sabaragamuwa Province, Ratnapura District, Halwathura-Kanda, near 
Rassagala, elev. 850 m, 15 Mar. 1978, M. E. Hale 51 216 (US—holotypus; BM—isotypus). 


Thallus greenish mineral gray, c. 10 cm broad, shiny, continuous; cortex 20-30 um, with 
aculeate hyphae, splitting and exfoliating; algal layer continuous, 15 ~m; medulla to 30 um, 
with crystals; apothecia immersed in the periderm, 0°3-0°4 mm diam, slightly raised at 
maturity, the exciple carbonized; columella c. 60 ~m diam; pore round, 0°! mm diam, white 
rimmed; hymenium 120-150 wm; spores brown, transversely septate, 6-7 x 14-18 wm, 5-6 
loculate, I—. 


CHEMISTRY. Psoromic acid. 
HaBiTAT. Trees along trail at mid elevations (850 m). 
DISTRIBUTION. Sri Lanka. 


OBSERVATIONS. A rather inconspicuous species, O. rassagala can be distinguished from 
other brown-spored members of Ocellularia because of the immersed columellate apothecia 
and chemistry. 


36. Ocellularia rhicnopora Hale sp. nov. 
(Fig. 20a) 


Thallus corticola, epiphloeodes, verrucosus vel fere granulosus, albido-cinereus, 15 cm latus; apothecia 
numerosa, irregulariter dispersa, semi-emergentia, amphithecio pallide brunneo, 0°7-1:°0 mm 
diametro, excipulo connato, columella nulla; ostiolum 0-1-0'2 mm diametro, nigro-cinctum; 
hymenium c. 904m altum; sporae incolores, transversim septatae, octonae, 8 x30 um, 7-9 
loculatae, I+ coeruleae. 

Typus: Sri Lanka, Southern Province, Galle District, Kanneliya Forest Reserve, elev. 
150 m, 16 Feb. 1976, M. E. Hale 46 147 (US—holotypus: BM—isotypus). 


Thallus pale yellowish brown to whitish, c. 15 cm broad, finely verruculose to subgranular; 
cortex loosely organized, 20-24 wm; algal layer continuous, 15 wm; medulla to 20 wm but 
mostly hypophloeodal; apothecia irregularly dispersed, semi-emergent, 0°7-1:0 mm diam, 
the exciple carbonized; columella lacking; pore irregular, 0° 1-0-2 mm diam, the surrounding 
area pale brown, rugose; hymenium c. 90 wm; spores colourless, transversely septate, 
8 x 30 wm, 7-9 loculate, I+. 


CHEMISTRY. No substances present. 

HABITAT. Lower trunk in rain forest at low elevations (150 m). 
DISTRIBUTION. Sri Lanka. 

ADDITIONAL SPECIMENS. Hale collections: 5 (46 246, 46 254). 


OBSERVATIONS. This species is apparently near the hypoprotocetraric acid-containing, 
ecolumellate, O. domingensis group, which in Sri Lanka includes O. kanneliyensis and O. 
neopertusariiformis. They all have large, emergent apothecia with a carbonized exciple, a 
rather loosely organized cellular cortex, and no aculeate hyphae or splitting. Ocellularia 
rhicnopora has an unusually broad, wrinkled pore area set off from the main thallus by the 
pale brown colour. 


THELOTREMATACEAE IN SRI LANKA 323 


2 & * - 
™ oe © jie 
i po 


Fig. 20 Species of Ocellularia. (a) O. rhicnopora (Hale 46 147). (b) O. sticticans (Hale 46 254). (c) 
O. subsimilis (Hale 51 141). (d) O. tenuis (Hale 46 267). (e) O. thelotremoides (Hale 51 040). (f) 
O. triglyphica (Hale 51 194). See Fig. 7i for scale. 


eos 


37. Ocellularia sticticans Hale sp. nov. 
(Fig. 20b) 
Thallus corticola, epiphloeodes, continuus, nitidus, 15 cm latus; apothecia vix emergentia, 0°4-0°7 mm 
diametro, excipulo fuligineo, columella 45-50 wm diametro; hymenium 35-40 um altum; sporae 


incolores, transversim septatae, octonae, 5-8 x 12-30 um, 6-8 loculate, I+ coeruleae. 
Typus: Sri Lanka, Southern Province, Matara District, Enselwatta, elev. 1100 m, 14 Feb. 1976, /. 


E. Hale 46 254 (US—holotypus; BM—isotypus). 
Thallus ashy white, c. 15 cm broad, thin but continuous, shiny; cortex dense, 12-20 um; 
algal layer continuous, 10-15 um; medulla to 30um; apothecia weakly emergent, 
0-4-0'7 mm diam, the exciple carbonized; columella 40-50 um, protruding into the pore; 
hymenium 35-40 wm; spores colourless, transversely septate, 5-8 x 12-30 wm, 6-8 loculate, 
I+. 


CHEMISTRY. Stictic acid. 
HABITAT. Lower trunks and sapling in rain forest at low to mid elevations (300-1100 m). 


. 


DISTRIBUTION. Sri Lanka, Solomon Islands. 


324 M. E. HALE 


ADDITIONAL SPECIMENS. Hale collections: 5 (46 246), 7a (SO 158). Other collections: 
Solomon Islands: Kolombangara Island, Hill 10 654 (BM, US); Santa Isabel Island, Hill 11 
249 (BM, US). 


OBSERVATIONS. One would probably identify this species as O. perforata without using a 
chemical test. Aside from the chemical difference, it has more consistently emergent 
apothecia. 


38. Ocellularia subsimilis (Hale) Hale in Mycotaxon 11 : 138 (1980). 
(Fig. 20c) 


Thelotrema subsimile Hale in Phytologia 27 : 497 (1974). Type: Trinidad, Arima, 17 April 1972, M. E. 
Hale 37 447 (US—holotype). 
Icongs. Hale, 19746 : 501 fig 17. 


Thallus whitish gray, 6-10 cm broad, shiny, smooth, continuous; cortex dense, 8-15 wm, 
with aculeate hyphae and some internal splitting; algal layer continuous, 10-15 wm; medulla 
0-60 um, with some crystals; apothecia immersed to slightly raised, 0-S-O°8 mm diam, the 
exciple carbonized; columella c. 60 ~m diam; pore round, 0:05—0:1 mm diam, the tip of the 
columella usually visible within; hymenium 120-150 um; spores colourless, muriform, 
6-10 x 22-30 um, 1-3 x 5-6 loculate, I+. 


CHEMISTRY. Protocetraric acid. 


HaBiTAT. On trees along trails occurring into the canopy branches in rain forest in a broad 
range of elevations (150-2100 m). 


DISTRIBUTION. West Indies, India, Sri Lanka, Philippines. 


ADDITIONAL SPECIMENS. Hale collections: 2 (51 141), 9 (46 134), 11 (50 317). Thwaites 
collections: C. L. 118 (BM, H-Nyl. 3866, PDA, UPS) (as ‘Lgt. 80. T. calvescens’). 


OBSERVATIONS. This species is externally identical with O. perforata, the main difference 
being the muriform spores. Leighton’s determination of C. L. 118 as ‘Thelotrema calvescens’ 
(= Myriotrema calvescens (Fée) Hale) may have been suggested and was at least confirmed by 
Nylander. Fée’s type contains psoromic acid and has transversely septate spores (see Hale, 
1978a : 15 and Salisbury, 1978 : 416). 


39. Ocellularia tenuis (Hale) Hale in Mycotaxon 11 : 138 (1980). 
(Fig. 20d) 


Thelotrema tenue Hale in Smithson. Contr. Bot. 16 : 38 (1974). Type: Dominica, Morne Diablotin, 
Jan. 1969, M. E. Hale 35 430 (US—holotype). 
Icongs. Hale, 1974a : 37 fig. 17k (holotype).—Hale, 1978a : 46 fig. 13). 


Thallus whitish mineral gray, 3-6 cm broad, shiny, thin, continuous; cortex lacking or very 
weakly developed as a thin polysaccharide layer with aculeate hyphae; algal layer 10 wm; 
medulla mostly hypophloeodal; apothecia immersed, 0°2-0'3 mm diam, the exciple weakly 
carbonized; columella 30-50 wm diam, sometimes lacking; pore round, 0°05 mm diam, 
becoming black rimmed; hymenium 70-90 wm; spores colourless, muriform 8-10 x 26- 
28 um, 2 x 6-7 loculate, I—. 


CHEMISTRY. No substances present. 

HABITAT. Trees along trail in mossy forest at mid elevations (1100 m). 
DISTRIBUTION. West Indies, Central and South America, Sri Lanka. 
ADDITIONAL SPECIMENS. Hale collections: 5 (46 267). 


OBSERVATIONS. Ocellularia tenuis is obviously related to O. melanotremata, which contains 
stictic acid (see above), because of the black-rimmed pore and ecorticate thallus. 


THELOTREMATACEAE IN SRI LANKA 325 


40. Ocellularia thelotremoides (Leighton) Zahlbr., Cat. Lich. Univers. 2 : 603 (1923). 
(Fig. 20e) 


Ascidium thelotremoides Leighton in Trans. Linn. Soc. Lond. 27 : 170 (1870). Type: Sri Lanka, Central 
Province, Thwaites C. L. 142 (BM—lectotype; G, H-Nyl. 3920, PC, S, UPS, W—isolectotypes) (as 
*Lgt. 97. Ascidium thelotremoides’) (C. L. 256 so determined in PDA is a pyrenocarpous lichen). 


Thallus pale greenish to straw coloured, 8-15 cm broad, shiny, continuous, slightly rugulose 
and white spotted with age; cortex dense, rather thin, 8-10 um; algal layer continuous, 
12-15 um; medulla 10-45 um, with crystals; apothecia immersed to slightly raised, 
0:3-0°6 mm diam, the exciple carbonized; columella c. 90 wm diam or sometimes lacking; 
pore round, 0°'1-0'2 mm diam; hymenium 95-150 um; spores colourless, muriform, 
6-10 x 18-30 um, 1-2 x 5-9 loculate, I+. 


CHEMISTRY. Fumarprotocetraric and protocetraric acids. 


HaBitTAT. Exposed roots, saplings, and base and lower trunk of trees in rain forest at low to 
mid elevations (150-850 m). 


DISTRIBUTION. India, Sri Lanka, Philippines. 


ADDITIONAL SPECIMENS. Hale collections: 1 (50 194, 50 196, 50 245, 50 259), 2 (51 115, 51 
174, 51 186, 51 198), 3b (46 388), 4 (50 379, 50 403, 50 423, 50 501, 50 502), 7b (51 002, 51 
019, 51 040, 51 074, 51 076), 8 (47 016, 47 030, 47 049, 47 051). 


OBSERVATIONS. This species often occurs with O. papillata and O. perforata, both of which 
resemble it externally but have transversely septate spores. I confirmed the chemistry of all 
collections with both hexane-ether-formic acid and n-butanol-acetone-water solvent 
systems. 


41. Ocellularia triglyphica (Krempelh.) v. Overeem & D. v. Overeem in Bull. Jard. bot. 
Buitenz. Ill, 4 : 119 (1922). 
(Fig. 20f) 


Ascidium triglyphicum Krempelh. in Nuovo G. bot. ital. 7 : 22 (1875). Type: Sarawak, Beccari 129b 
(M—lectotype). 

Ascidium melanostomum Krempelh. in Flora, Jena 59: 248 (1876). Type: Brazil, Prov. Rio de 
Janeiro, Glaziou 5573 (M—lectotype; BM, G, PC, W—isolectotypes). 

Ascidium majorinum Nyl., Sert. Lich. Trop. Labuan Singapore : 20 (1891). Type: Singapore, 1879, 
Almquist s.n. (H-Nyl. 22 387—lectotype; S—isolectotype). 

Phaeotrema melanostomum (Krempelh.) Mill. Arg. in J. Linn. Soc. (Bot.) 30 : 454 (1895). 

Ocellularia majorina (Nyl.) Zahlbr., Cat. Lich. Univers. 2 : 595 (1923). 

Thallus pale greenish mineral gray, S—12 cm broad, shiny, smooth to verruculose with age; 

cortex cellular, 10-15 wm; algal layer 10 wm; medulla 5-60 wm, with crystals, mostly 

hypophloeodal; apothecia conspicuous, strongly emergent, the base sometimes constricted, 

0:8-1:1 mm diam, the exciple carbonized; columella 120-160 um; pore round to irregular, 

0:1-0:2 mm diam; white rimmed; hymenium 160-180 wm; spores colourless, transversely 

septate, 10-15 x 70-150 um, 18-20 loculate, I+. 


CHEMISTRY. Hypoprotocetraric acid, 4-O-demethylnotatic acid, and unknowns GU | and 2 
(fide C. F. Culberson). 


HaBitaT. Base and lower trunks of trees in rain forest at mid or lower elevations 
(300-850 m). 


DISTRIBUTION. South America, Sri Lanka, Peninsular Malaysia, Philippines, Sarawak. 


ADDITIONAL SPECIMENS. Hale collections: 2 (51 187, 51 194, 51 202); 4 (50 411, 50 458, 50 
467, 50 478), 7b (51 004). 


326 M. E. HALE 


OBSERVATIONS. This conspicuous lichen has an extensive synonymy. Dr C. F. Culberson 
kindly established the identity of the chemistry in the types of Ascidium triglyphicum and A. 
majorinum. Other columellate hypoprotocetraric acid-containing species on the world level 
include O. phaeotropa (Krempelh.) Mill. Arg., which has small spores (6 x 19 wm), and two 
large spored species, O. granulifera (Krempelh.) Miill Arg. with a coarsely granular thallus 
and O. henatomma (Ach.) Mull. Arg. with a smoother thallus and small pore (see Hale, 
1972 : 192). Among the muriform spored species we can find O. arecae (Vainio) Hale, which 
has the same chemistry as O. triglyphica, and O. grandis (Hale) Hale, which has a strongly 
annulate pore and very large spores. The chemistry of these species is discussed in more 
detail by Culberson & Hale (1973). 

Redinger (1936 : 26) considered Ascidium melanostomum to be a synonym of Ocellularia 
rhabdospora (Nyl.) Redinger, but that species, known only from the New World, contains 
the P+ red ‘cinchonarum’ unknown (Hale, 1978a: 29). 


Acknowledgements 


I would like to offer thanks to curators who sent loans of type specimens as cited in the text. I am 
especially grateful to Mr Peter W. James of the British Museum (Natural History) for providing work- 
ing facilities and help with literature during my many visits. Mr Orvo Vitikainen was indispensable in 
finding the Leighton correspondence at the University library in Helsinki. 

The field work in Sri Lanka was aided immeasurably by the Smithsonian Flora of Ceylon Project, 
directed by Dr F. Raymond Fosberg, which provided transportation. Mr Shelton Waas was very helpful 
during my trip in 1978; it was he who located most of the logging sites and assisted in the field. 


References 


Alston, A. H. G. 1938. The Kandy Flora. Colombo. 

Berkeley, M. J. & Broome, C. E. 1875. Enumeration of the fungi of Ceylon. J. Linn. Soc. (Bot.) 
14 : 29-140. 

Brenan, J. P. M. & Ross, R. 1970. Reorganization of the cryptogamic collections at the Royal Botanic 
Gardens, Kew and the British Museum (Natural History). Lichenologist 4 : 257.. 

Crabbe, J. A. 1960. A. H. G. Alston (1902-1958). A bibliography of his writings, with a short intro- 
duction and a list of new taxa and nomenclature changes published by him. J. Soc. Biblphy nat. Hist. 
3 : 383-404. 

Crombie, J. M. 1891. Recent observations of Dr Nylander on Schwendenerism. Grevillea 20 : 60-62. 
[Translation from Nylander.] 

Culberson, C. F. & Hale, M. E., Jr. 1973. 4-O-DemethyInotatic acid, a new depsidone in some lichens 
producing hypoprotocetraric acid. Bryologist 76 : 77-84. 

Fries, T. M. 1871. Lichenographia Scandinavica sive dispositio lichenum in Dania, Suecia, Norvegia, 
Fennia, Lapponia Rossica Lactenus collectorum. Upsalia. 

Gilenstam, G. 1969. Studies in the lichen genus Conotrema. Ark. Bot. II, 7(2) : 149-179. 

Hale, M. E., Jr. 1972. Typification of species in the lichen family Thelotremataceae described by 
Acharius. Bot. Notiser 125 : 186-198. 

—— 1973a. Fine structure of the cortex in the lichen family Parmeliaceae viewed with the scanning- 
electron microscope. Smithson. Contr. Bot. 10: 1-92. 

— 1973b. Studies on the lichen family Thelotremataceae. |. Phytologia 26 : 413-420. 

— 1974a. Morden-Smithsonian expedition to Dominica: The lichens (Thelotremataceae). Smithson. 
Contr. Bot. 16: 146. 

— 1974b. Studies on the lichen family Thelotremataceae. 2. Phytologia 27 : 490-501. 

—— 1975. Studies on the lichen family Thelotremataceae. 3. Mycotaxon 3 : 173-181. 

— 1978a. A revision of the lichen family Thelotremataceae in Panama. Smithson. Contr. Bot. 
38 : 1-60. 

—— 1978b. Studies on the lichen family Thelotremataceae. 4. Mycotaxon 7 : 377-385. 

— 1980. Generic delimitation in the lichen family Thelotremataceae. Mycotaxon 11 : 130-138. 

Harmand, J. 1912. Lichens recueilles dans la Nouvelle-Calédonie ou en Australia par la R. P. 
Pionnier, missoinaire et determinés par M. l’abbé J. Harmand. Bull Séanc. Soc. Sci. Nancy Ill, 13: 
36-64. 


THELOTREMATACEAE IN SRI LANKA 327) 


Henssen, A. & Jahns, H. M. [‘1974’]. Lichenes. Eine Einfiihrung in die Flechtenkunde. Stuttgart. 

Krempelhuber, A. 1877. Die Flechten-Gattung Ascidium Fee. Ber. bot. Ver. Landshut 1877 : 119-138. 

Leighton, W. A. 1851. The British Species of Angiocarpous Lichens, Elucidated by their Sporidia. 
London. 

— 1854. A monograph of the British Graphideae. Ann. Mag. nat Hist. Il, 13 : 81-97, 203-212, 
265-279, 387-395, 437-446. 

— 1870. The lichens of Ceylon collected by G. H. K. Thwaites, Esq., Ph.D., F.R.S., F.L.S., Director 
of Royal Botanic Garden, Peradeniya, Ceylon. Examined and determined by the Rev. W. A. 
Leighton, B.A., F.L.S., F.B.S. Edin. Trans. Linn. Soc. Lond. 27 : 161-185. 

—— 1871. The Lichen-flora of Great Britain, Ireland and the Channel Islands. Shrewsbury. 

Letrouit-Galinou, M. A. 1966. Recherches sur l’ontogenie et l’'anatomie comparées des apothécies de 
quelques Discolichens. Revue bryol. lichén. 34 : 413-588. 

Magnusson, A. H. 1939. Studies in species of Lecanora mainly the Aspicilia gibbosa group. K. svenska 
VetenskAkad. Handl. Ill, 17(5) : 1-182. 

— & Zahlbruckner, A. 1944. Hawaiian lichens. Ark. Bot. 31A(1): 1-96. 

Montagne, J. F.C. 1856. Sylloge Generum Specierumque Cryptogamarum. Paris. 

Miller [Argoviensis], J. 1887. Graphideae Feeane. Mém. Soc. Phys. Hist. nat. Genéve 29(8) : 1-80. 

Nordenskiéld, A. E. 1881. Vegas Fard Kring Asien och Europa jemte en historisk Aterblick pa 
foregaende Resor langs gamla Verldens Nordkust (Senare delen). Stockholm. 

Nylander, W. 1862. Conspectus generis Thelotrematis. Annis Sci. nat. (Bot.) 1V, 16 : 95-96. 

— 1863. Lichenographiae Novo-Granatensis Prodromus. Acta Soc. Scient. fenn. 7 : 417-504. 

— 1866. Hypochlorite of lime and hydrate of potash, two new criteria in the study of lichens (transl. 
by W. A. Leighton). J. Linn. Soc. (Bot.)9 : 358-365. 

— 1869. Lichenes Kurziani Bengalienses collectio altera. Flora, Jena 52 : 69-73. 

— 1891. Lichenes Pyrenaeorum Orientalium Observatis Novis. Paris. 

— 1900. Lichenes Ceylonenses et additamentum ad lichenes Japoniae. Acta Soc. Scient. fenn. 
26(10) : 3-33. 

Patwardhan, P. & Kulkarni, C. 1977. Some new taxa of the family Thelotremataceae from Western 
Ghats, SW India. Norw. J. Bot. 24: 127-131. 

Phillips, W. 1889. William Allport Leighton. J. Bot., Lond. 27: 111-113. 

Redinger, K. M. 1936. Thelotremaceae Brasilienses imprimis ex herbario Regnelliano cognitae 
praetereaque in herbariis Krempelhuberi, Miilleri Arg., Nylanderi, Wainionis et Zahlbruckneri 
asservatae. Ark. Bot. 28A(8) : 1-122. 

Salisbury, G. 1971la. Thelotrema wightii (T. Tayl.) Nyl. Port. Acta biol. 11 : 5-7. 

— 1971b. The Thelotremata of Angola and Mocambique. Revita Biol. Lisb. 7 : 271-280. 

— 1972a. Thelotrema sect. Thelotrema |. The T. lepadinum group. Lichenologist 5 : 262-274. 

—1972b. Thelotrema) sect. Thelotrema 2. The T. platycarpum group. Revue bryol. lichén. 
38 : 281-290. 

— 1975. Thelotrema monosporum Nyl. in Britain. Lichenologist 7 : 59-61. 

— 1978. Thelotremata Achariana et Féeana. Nova Hedwigia 29 : 405-427. 

Santesson, R. 1952. Folticolous lichens I. A revision of the taxonomy of the obligately folticolous, 
lichenized fungi. Symb. bot. upsal. 12(1) : 1-590. 

Sherwood, M. A. 1977. The Ostropalean fungi. Mycotaxon 5 : 1-277. 

Thwaites, G. H. K. 1864. Enumeratio Plantarum Zeylaniae: an Enumeration of Ceylon Plants, with 
Descriptions of the New and Little-known Genera and Species, Observations on their Habitats, Uses, 
Native Names, etc. London. 

Tuckerman, E. 1864. Observationes lichenologicae. Observations on North American and other 
Lichenes. Proc. Am. Acad. Arts Sci. 6 : 263-287. 

Vainio, E. A. 1921. Lichenes Insularum Philippinarum, III. Swomal. Tiedeakat. Toim. A, 
15(6) : 1-368. 

Vézda, A. 1966. Flechtensystematische Studien III. Die Gattungen Ramonia Stiz. und Gloeolecta Lett. 
Folia geobot. phytotax. Bohemoslovaca 1 : 154-175. 

Wirth, M. & Hale, M. E., Jr. 1978. Morden-Smithsonian expedition to Dominica: The lichens 
(Graphidaceae). Smithson. Contr. Bot. 40 : 1-64. 

Zahlbruckner, A. 1909. Lichenes (Flechten). Jn R. Wettstein and V. Schiffner, Ergebnisse der 
botanische Expedition der Akademie der Wissenschaften nach Stidbrasilien 1901. Denkschr. Akad. 
Wiss. Wien (Math.-nat. KI.) 83 : 87-211. 

— 1923. Catalogus Lichenum Universalis. 2. Leipzig. 

— 1941. Lichens Novae Zelandiae a cl. H. H. Allan eiusque collaboratoribus lecti. Denkschr. Akad. 
Wiss. Wien (Math.-nat. KI.) 104: 1-132. 


Accepted names are given in roman, whilst new names are in bold, as are principal references. An 


asterisk (*) denotes a figure. 
Anthracothecium 242, 278 


desquamans 236, 278 
Antrocarpum 240, 249 


Ascidium 230, 240, 241, 242, 297 


depressum 314, 315 
granulosum 230, 283 
majorinum 325, 326 
var. longius 316 
massalongoi 311 
melanostomum 325, 326 
monobacterium 311 
orthomastium 316 
pachystomum 230, 311 
pertusariiforme 230 
punctulatum 230, 320, 321 
thelotremoides 230, 325 
triglyphicum 325, 326 
Asteristion 239, 249 
erumpens 230, 231,265 


Brassia 249 


Chroodiscus 239, 249 
coccineus 253 

Conotrema 242 

Coscinedia 271 


Diploschistes 242 


Ectolechia 297 
Endocarpon wightii 296 


Gloeslecta 242 
Gomphospora 249 
Graphis alborosella 251 
Dhlyctidea 258 
subnivescens 258, 259 
Gyrostomum 242 


Leptotrema 240, 242, 276, 297 
andamanicum 275 
compunctum 276 
crassum 241, 302 
desquamans 278 
diffractum 281 
elachistoteron 278 
fallax 302 
fissum 307 
glaucescens 282 
inclusum 28 | 
integrum 302 
irosinum 278 
marivelense 310 
microglaenoides 279 
monosporoides 314 


monosporum 241, 260, 279 
oleosum 307 

polycarpum 281 
polyporum 281 
sandwicense 307 
subgeminum 279 
vesiculiferum 302 

wightii 296 


Macropyrenium 297 


pertusarioides 311 


Myriotrema 233, 238, 239, 240, 241, 242, 


244, 245, 246, 247, 248, 249, 264, 271, 
278, 282, 286, 292, 302 

albocinctum 227, 235, 243, 244, 245, 246, 
DN 2S 24s 

album 235*, 239*, 243, 244, 246, 271, 272, 
2735.2 74*, 275, 276, 287, 29129282948 
295, 296 

anamalaiense 233, 243, 246, 271, 272, 274*, 
275 

andamanicum 233, 237, 239, 243, 271, 274*, 
275, 286 

calvescens 324 

cinereoglaucescens 235, 243, 246, 271, 272, 
274*, 276, 292 

cinereum 288 

clandestinum 235, 293 

compunctum 238, 243, 246, 271, 273, 274*, 
276, 277, 292, 321 

concretum 284, 293 

conforme 294 

costaricense 235, 243, 246, 271, 272, 274*, 
277 

deceptum 278 

decorticatum 227, 233, 243, 246, 271, 273, 
Daa 

desquamans 253, 243, 244, 245, 246, 271, 
273, 274*, 278 

elachistoteron 234, 243, 246, 271, 272, 278, 
279, 280* 

eminens 235, 243, 244, 246, 271, 272, 279, 
280* 

fissurinum 227, 235, 243, 246, 271, 273, 278, 
279, 280* 281, 289 

fluorescens 227, 235, 236*, 243, 244, 246, 
271, 272, 280*, 281, 282 

frondosum 227, 235, 243, 246, 271, 272, 
280*, 282 

glaucescens 237*, 238, 243, 246, 271, 273, 
280*, 282 

glaucophaenum 235, 236*, 243, 246, 271, 
280*, 283 

glauculum 287 


THELOTREMATACEAE IN SRI LANKA 329 


granulosum 234, 243, 244, 246, 271, 280*, 
283 

hartii 235, 243, 246,271, 272, 284, 285* 

immersum 309 

masonhalei 235, 243, 245, 246, 271, 272, 284, 
285*, 288 

mastarion 227, 233, 237, 239, 243, 246, 271, 
272, 276, 285*, 286 

microporellum 295 

microporum 234, 237*, 238, 243, 244, 246, 
271, 272, 275, 285*, 286, 287, 289, 291, 
295 

microstomum 235, 243, 244, 245, 246, 271, 
DD 285*) 287 

minutulum 235, 243, 246, 271, 272, 273, 
285*, 288 

minutum 239*, 243, 246, 271, 272, 285*, 288 

multicavum 227, 233*, 237*, 238, 243, 246, 
271,272, 285*, 288 

myrioporum 275 

norsticticum 292 

nuwarense 227, 234, 243, 244, 246, 271, 272, 
289, 290* 

occultum 292 

olivaceum 235, 241, 243, 244, 245, 246, 271, 
272, 275, 287, 289, 290*, 291, 295 

phaeosporum 278 

polytretum 227, 235, 244, 246, 271, 272, 
290*, 291 

porinaceum 233, 243, 246, 271, 272, 290*, 
292 

protoalbum 227, 235, 243, 246, 271, 272, 
290*, 292 

reclusum 278, 292 

rugiferum 243, 245, 246, 271, 272, 287, 290*, 
292, 293 

santessonii 235, 243, 246, 271, 272, 290*, 293 

secernendum 288 

subcompunctum 278, 281 

subconforme 235, 243, 246, 271, 272, 290*, 
294 

terebrans 233*, 235, 243, 246, 271, 272, 290*, 
294 

terebratulum 235, 243, 244, 246, 271, 272, 
275, 283, 287, 289, 291, 294, 296* 

thwaitesii 227, 233, 244, 246, 271, 272, 287, 
295, 296* 

trypaneoides 278 

uniseptatum 275 

viridi-album 277 

wightii 238, 243, 246, 271, 273, 296* 


Ocellularia 239, 240, 242, 245, 247, 248, 249, 
Pile 297.322 

alba 273,275,277 
f. costaricensis 277 

albidiformis 266 

albomaculata 227, 235, 243, 246, 298, 299*, 
300 


alborosella 251 

antillensis 309 

arecae 305, 326 

ascidioidea 227, 233, 243, 246, 298, 299*, 
300 

asiatica 320 

astroidea 253 

auberiana 251, 310 

aurata 234, 243, 244, 246, 298, 299* 300 

cavata 315, 321 

chonestoma 230, 231, 233, 235*, 243, 244, 
245, 246, 298, 299*, 301, 316 

cinereoglaucescens 276 

coccinea 253 

comparabilis 309 

conferta 302 

conglomerata 262 

costaricensis 277 

crassa 235*, 243, 244, 245, 246, 298, 299*, 
302, 307, 311, 320 

crassula 286 

cricota 256 

crocea 315 

croceopora 227, 235, 236*, 243, 244, 246, 
297, 299*, 302, 309 

decolorata 315 

diacida 234, 243, 245, 246, 298, 299*, 303 

dilatata 254 

discoidea 309 

dodecamera 315 

dolichospora 256 

dolichotata 233, 235, 243, 244, 246, 297, 
299*, 303 

domingensis 305, 308, 322 

efformata 304 

emersa 235, 236*, 241, 243, 246, 298, 299*, 
304, 305 

emersella 304 

erumpens 314 

eumorpha 233, 243, 246, 298, 305, 306*, 


308 
exanthismocarpa 267 


exigua 315 

exuta 227, 234, 243, 244, 246, 298, 305, 306* 

fissa 231, 235, 243, 246, 282, 298, 302, 304, 

305, 306*, 307, 308, 311 

galactina 294 

glaucophaena 283 

grandis 326 

granulifera 326 

henatomma 326 

interponenda 313 

interposita 305 

Japonica 256 

kamatii 255 

kanneliyensis 227, 234*, 243, 246, 298, 306*, 
308, 322 

keralensis 227, 234, 243, 244, 245, 246, 298, 
306*, 308, 309 


330 M. E. HALE 


lankaensis 227, 235, 237*, 243, 244, 245, 
246, 297, 298, 303, 306*, 309 

leightonii 283 

leucina 320 

leucomelaena 232, 243, 245, 246, 298, 306*, 
309 

lirelliformis 234, 243, 246, 298, 306*, 310 

lopezii 266 

magnifica 259 

majorina 325 

marivelensis 236, 243, 244, 246, 298, 310, 
312* 

massalongoi 236, 243, 244, 246, 298, 311, 
S12F 53113 

meiosperma 232, 243, 246, 298, 312*, 313, 
314 

meiospermoides 227, 232, 243, 246, 298, 
312*, 313 

melanotremata 227, 232, 243, 244, 246, 298, 
312*, 314, 324 

micropora 286 

minuta 288 

minutula 288 

monosporoides 234, 240, 243, 244, 246, 298, 
312*, 314, 315 

neocavata 227, 234, 243, 245, 246, 298, 312*, 
315 

neodominicana 315 

neopertusariiformis 227, 233, 243, 246, 297, 
308, 312*, 315, 316, 319, 322 

nylanderiana 233, 243, 244, 245, 246, 297, 
312*, 316 

obovata 267 

olivacea 289 

orthomastia 233, 243, 246, 297, 316, 318* 

papillata 236, 243, 244, 245, 246, 297, 300, 
Siig Bless SHG WS) 

patwardhanii 240 

perforata 234*, 235, 243, 244, 245, 246, 298, 
303, 317, 318*, 319, 324, 325 

pertusariiformis 234, 235, 243, 246, 297, 316, 
318*, 319 

phaeotropa 326 

phlyctidioides 262 

platystomum 305 

pluripora 227, 236, 243, 246, 298, 318*, 319 

polillensis 227, 236, 243, 244, 245, 246, 298, 
318*, 320 

polydisca 315 

pseudoexanthismocarpa 267, 268 

punctulata 235, 243, 244, 245, 246, 297, 315, 
318*, 320, 321 

pyrenuloides 243, 246, 297, 318*, 321 

rhabdospora 315, 326 

rassagala 227, 236, 243, 245, 246, 298, 318*, 
322 

rhicnopora 227, 234, 243, 246, 297, 298, 305, 
B22, 623r 

rhodostroma 240 


sandfordiana 305 

siamensis 303 

sticticans 227, 234, 243, 246, 298, 323* 

Stictidea 231 

straminea 320 

subcavata 240 

subemersa 315 

subsimilis 236, 243, 246, 298, 323*, 324 

tenuis 232, 243, 245, 246, 298, 322, 323* 324 

terebrans 294 

terebrata 320 

terebratula 294 

thelotremoides 234, 243, 244, 245, 246, 298, 
323*, 325 

triglyphica 234, 243, 246, 297, 323*, 325, 326 

virens 305 

viridi-alba 277 

xanthostromiza 315 


Pertusaria 264 
Phaeographis phlyctidea 258 
subnivescens 258 
Phaeographina 269 
Phaeotrema 240, 242, 313 
aggregatum 251 
apertum 265 
cricotum 256 
emersum 304 
erumpens 265 
lacteum 241,256 
leprieurti 258 
lirelliforme 310 
melanostomum 325 
platycarpoides 265 
platycarpum 265 
rockii 258 
subinalbescens 313 
virens 241, 305 
Platygrapha 260 
albo-rufa 265 
astroidea 253 
bivela 265 
coccinea 253 
magnifica 259 
Polystroma 242 
Porophora 240, 297 
Pyrenula discissa 313 


Ramonia 242 

Rhabdodiscus 239, 240, 277, 297 
costaricensis 277 
fissus 307 
lirelliformis 310 


Stegobolus 239,297 
Stictis 242 
leprieurii 258 


THELOTREMATACEAE IN SRI LANKA 331 


Thelotrema 238, 239, 240, 241, 242, 244, 
245, 246, 247, 248, 249, 269 

aggregatum 234, 243, 246, 249, 250, 251, 
DiS2* 

albidiforme 229, 230, 231, 266, 267 

alborosellum 232, 243, 244, 246, 250, 251, 

PSP 253 

album 271, 273, 282 

anamalaiense 275 

andamanicum 275 

arecae 305 

astroideum 234, 243, 244, 246, 250, 252*, 
253 

auberianum 251,310 

auratum 300 

bahianum 242, 271 

calvescens 324 

cavatum 297 
var. confertum 301 
var. dolichosporum 256 

chonestomum 301 

cinereoglaucescens 276 

clandestinum 293 

coccineum 232, 243, 246, 249, 252*, 253 

colobicum 232, 243, 244, 246, 250, 252*, 253 

compunctum 281 

concretum 293 

conferendum 260 

confluens 259 

conforme 294 

crassulum 286 

cryptotrema 241, 264 

curranii 230 

deightonii 253 

depressum 315 

dilatatum 233, 234*, 243, 246, 250, 252*, 

254, 262,270 

diminitum 266 

disciforme 260 

discoideum 297 

discolor 297 

dislaceratum 264 

dissultum 227, 234, 243, 246, 250, 252*, 254 

dolichosporum 256 

dolichotatum 303, 304 

dominicanum 315 

efformatum 304 

elachistoteron 278 

emersum 304 

eminens 279 

epitrypum 307 

eumorphum 305 

exanthismocarpum 231, 266, 267 

fissum 307 

gibbosum 238, 264 

glaucescens 282 

glaucopallens 281 

glaucophaenum 283 

hartii 284 


hawaliense 260 
hians 313 
imperfectum 227, 234, 235*, 243, 245, 246, 
250) 252 2554250, 20 
interpositum 305 
irosinum 278 
kamatii 234, 243, 246, 249, 250, 252*, 255, 
261 
lacteum 233, 243, 246, 248, 250, 256, 257*, 
260 
leightonii 311 
lepadinum 233, 238, 241, 243, 244, 246, 249, 
250, 257*, 258, 260, 264 
*lepadizum 232,258 
lepadodes 260 
leprieurii 234, 243, 246, 250, 257*, 258 
leprocarpoides 227, 233, 243, 246, 250, 257*, 
259 
leprocarpum 254 
leptoporum 287 
leucomelaenum 309 
leucophthalmum 254 
lirelliforme 310 
magnificum 236, 243, 244, 245, 246, 249, 
254, 257*, 259 
marginans 311 
marivalense 310 
masonhalei 284 
massalongoi 311 
meiospermum 313 
microporellum 295 
microporum 286, 287 
microstomum 287 
var. formosanum 287 
mirabile 269 
monobactrium 311 
var. endoleucum 311 
var. endorhoda 311 
var. geminiparum 31 | 
var. subgranulatum 311 
monosporoides 314 
monosporum 233, 240, 243, 246, 248, 250, 
ASG), PO 5 2 Nee AND, 2S), 3S) 
myrioporum 275 
nostalgicum 233, 243, 244, 246, 249, 257*, 
260, 261, 264,271 
nova-zelandiae 233, 243, 246, 250, 257*, 261 
nureliyum 227, 234, 239*, 243, 244, 246, 250, 
ASSP )5 2S AAO 
obovatum 267 
obturatum 297 
occlusum 305 
olivaceum 287, 309 
pachystomum 311 
papillatum 317 
patens 233, 243, 246, 250, 254, 262, 263* 
perforatum 317 
periphysatum 238 
pertusariaeforme 230, 319 


332 M. E. HALE 


phaeospermum 313 schizostomoides 307 
phaeosporum 278 schizostomum 304 
phlyctidioides 233, 243, 246, 250, 262, 263* siamense 303, 304 
pidurutalagalum 227, 240, 243, 244, 246, 250, sordidescens 321 
26,262, 263* 271 sphinctrinellum 229, 241, 264, 266 
piluliferum 236, 238, 241, 243, 244, 246, 249, stellatum 269 
250, 263*, 264, 266 subconcretum 296 
platycarpoides 234, 243, 246, 250, 263*, 265, subconforme 294 
266 subinalbescens 313 
platycarpum 231, 234, 241, 243, 246, 249, subpatens 227, 233, 239*, 243, 246, 250, 
250, 251,254, 259, 262, 263* 265, 266 268*, 269 
platysporum 236, 241, 244, 245, 246, 249, subsimile 324 
250, 263*, 264, 266, 277, 307 subtile 255, 256, 261 
platystomum 304, 305 tenue 324 
plurifarium 295 terebrans 294 
polillense 320 terebratulum 294 
porinaceum 292 turgidulum 305 
porinoides 229, 230, 231, 233, 243, 244, 245, vernicosum 264 
246, 250, 260, 262, 263*, 266, 267, 268 vesiculiferum 302 
pseudoexanthismocarpum 234, 243, 246, viridi-album 277 
250, 267, 268*, 269 waasii 227, 233, 234*, 243, 246, 249, 254, 
pulvereodiscum 227, 234, 243, 246, 250, 268 268*, 270 
punctulatum 229, 277, 321 weberi 233, 243, 244, 246, 250, 261, 264, 
pycnophragmium 270 268*, 270 
reclusum 292 Tremotylium 242 
rockii 258, 260 angolense 242 
rugiferum 292 sprucei 242 
sandfordianum 305 Trypethelium schizostomum 282, 307 
santessonii 293 
saxicola 241, 264 Urceolaria compuncta 276 


scabiomarginatum 227, 235, 243, 244, 245, 
246, 249, 250, 268*, 269 Verrucaria discissa 313 


> 


= 


British Museum (Natural History) 
1881-1981 


Centenary Publications 


Chance, change & challenge 


Two multi-author volumes from one of the foremost scientific institutions in the world. 


General Editor: P. H. Greenwood 


The Evolving Earth 


Editor: L. R. M. Cocks 


The Evolving Biosphere 
Editor: P. L. Forey 


In the first volume, The Evolving Earth, twenty scientists have been asked to review 
the present state of knowledge in their particular field, ranging from the origin of 
the Earth, through ocean sediments and soils to continental drift and palaeogeography. 


In the companion volume, The Evolving Biosphere, museum scientists have chosen 
an evolutionary concept—speciation, coevolution, biogeography etc. and related 
this to the group of animals or plants in which they are specialising. Thus beetles 
and birds exemplify sympatric and alloptaric speciation, butterflies mimicry and 
certain fishes explosive evolution. 


In both volumes the text is supplemented by over one hundred specially-commissioned 
pieces of two-colour artwork. 


These two books will be invaluable to all sixth-form and undergraduate biology and 
geology students. 


The Evolving Earth: 276x219 mm, 280pp, 138 line illustrations, 42 halftones 
The Evolving Biosphere: 276219 mm, approx. 320pp, 133 line illustrations 
Publishing: Spring 1981 


Co-published by the British Museum (Natural History), London and Cambridge 
University Press, Cambridge. 


Titles to be published in Volume 8 eee E : 


The Thelypteridaceae of Ceylon. By W. A. Sledge =: . 
Studies in the genus Hypericum L. (Guttiferae) 
2. Characters of the genus. By N. K. B. Robson 
By Mason E. Hale a 


Vascular plant collections from the Tristan da Cintas group of 
islands. By Eric W. Groves 


Printed by Henry Ling Ltd, Dorchester 


Bulletin of the 
British Museum (Natural History) : 


_ Vascular plant collections from the 
_ Tristan da Cunha group of islands 


_ Eric W. Groves 


_ Botany series Vol8No4 30 July 1981 


The Bulletin of the British Museum (Natural History), instituted in 1949, is issued in four 
scientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology, a 
an Historical series. 


Papers in the Bulletin are primarily the results of research carried out on the unique and : oe is 
ever-growing collections of the Museum, both by the scientific staff of the Museum and by — <The 
specialists from elsewhere who make use of the Museum’s resources. Many of the papaa’ are 
works of reference that will remain indispensable for years to come. on 


Parts are published at irregular intervals as they become ready, each is complete in itself, _ 
available separately, and individually priced. Volumes contain about 300 pages and several 
volumes may appear within a calendar year. Sees may be eg o8 for one or more of 


the individual parts. Orders and enquiries should be sent to: ra 


Publications Sales, 
British Museum (Natural History), eee - 
Cromwell Road, 7H ~ 
London SW7 5BD, he: 3 
England. 


World List abbreviation: Bull. Br. Mus. nat. Hist. (Bot.) 
© Trustees of the British Museum (Natural History), 1981 


This number completes Volume 8 


ISSN 0068-2292 Botany series 


British Museum (Natural History) ss 
Cromwell Road ees - 
London SW7 5BD Issued 30 July 1981 : 


Vascular plant collections from the Tristan da 
Cunha group of islands 
Eric W. Groves 


Department of Botany, British Museum (Natural History), Cromwell Road, London 
SW7 5BD 


Contents 
USS MN MRM kk 333 
TIMOGUCHONM MMMM EAS CS. 0 Wt os we kl 333 
History andtopographical description . . . . . . . ..). 335 
Visitors who have made vascular plantcollections. . . . . .. . 342 
(NTONOLOPICAIINISH pe ee kl 342 
FP HADEULCAUNISUMEMEMENEPMN TS 9 hc se gw Sw 346 
Systematielist . = 4%: . . . : ‘ ; ox Gere ee 354 
SVPQTAMENOTDIIIDS. 3 ea ek se 354 
Pteridophyta . : ; ie : ; : : : : : : : 400 
Comments on the flora Pe i 410 
INGRKNOWMMCGEEMeMISE 9 es. 5s 2 wlll 414 
IRGATOMNEES se 415 
Index tospecies . a toe i. “os 9S a a an : 417 

Synopsis 


An account of the collections of flowering plants and ferns made on the Tristan da Cunha group of 
islands is presented. A brief historical and topographical summary of each island is followed with a 
chronological and an alphabetical list of persons, expeditions and ships that have visited and collected 
vascular plants on the group. The systematic list gives details of the 212 vascular plant taxa recorded, 
which consists of 177 spermatophyte taxa (including 58 natives, of which 34 are endemic, and 119 
aliens) and 35 pteridophyte taxa (all native including 20 endemics). Included are the descriptions of 
one new genus Parodiochiloa, ten new species (Agrostis goughensis, A. holgateana, A. trachychlaena, A. 
wacei, Calamagrostis deschampsiiformis, Deschampsia christophersenii, D. mejlandii, D. robusta, D. 
wacei, and Glyceria insularis) one new sub-species (Agrostis magellanica subsp. laeviuscula), and two 
new combinations (Parodiochloa flabellata and Polypogon mollis). All belong to the family Gramineae 
and are provided by C. E. Hubbard, with the exception of Polypogon mollis which is made in conjunc- 
tion with the present author. 


Introduction 


Under each of the 212 vascular plants enumerated is given the scientific name and a brief 
relevant synonymy, the island vernacular name (where known), status, first known record, 
and habitat details. These last are a synthesis of the notes from collectors’ labels. Inclusion of 
full data from every sheet would have made the text unnecessarily long, so data relating to all 
cited specimens have been recorded and stored in the data-bank associated with the British 
Antarctic Survey’s bryophyte herbarium housed at the Research Station, Institute of 
Terrestrial Ecology, Bush Estate, Penicuik, Midlothian, Scotland. 


Bull. Br. Mus. nat. Hist. (Bot.) 8 (4) : 333-420 Issued 30 July 1981 
333 


E. W. GROVES 


334 


“‘APIVY JISIPY IG Aq INO]OI19}eM |BUISIIO UR WOI “976] Alenues (W 8¢Qz) 
Yea IY} SIDAOD PNO[D MOTT “ddURISIP WY € INOgP 1e Bas dy} WO POMaIA ‘eYUND ep UR\sII] [SI 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 335 


uSé Helena 


/ 


~ 
Tristan da Cunha 2900 


° 
0) a South Georgia ~ Bouveé 


Fig. 2 Map showing the position of the Tristan da Cunha islands in the south Atlantic and their 
relative distance (in km) from adjacent continents and islands. 


Following the habitat details is a citation of those herbarium specimens which have been 
examined by the author, indicated by collectors’ names and numbers, together with 
abbreviations of the herbaria where such material is at present located. The citation of these 
abbrevations are given in order of the size and importance of the sets held. Thus, the first 
herbarium mentioned has the major set that contains the specimen in question; those 
following hold smaller sets. Institution abbreviations follow Index Herbariorum (Holmgren 
& Keuken, 1974) (see also pages 346 & 414). 

This contribution represents the first listing of all available collections to date of vascular 
plants from the Tristan da Cunha group. Some may now possibly be extinct. Together with 
the earliest record for each species it forms a basic source work against which all future 
collections can be evaluated. Unfortunately I have been unable to make a personal visit to 
the islands. 


History and topographical description 


General 

The Tristan da Cunha islands (including Gough Island) are an isolated group situated in the 
middle of the South Atlantic Ocean approximately mid-way between South Africa 
and South America (Fig. 2). They lie approximately 2400 km south of St Helena and 
6440 km from the Antarctic continent. The ‘Tristan-Gough group’ as interpreted through- 
out this paper includes not only a cluster of five main islands: Tristan da Cunha, Inacces- 
sible, Nightingale, Middle and Stoltenhoff, but also Gough Island which lies approx. 352 km 
further to the south- south-east and forms a natural member of the group (Fig. 3). The 


336 E. W. GROVES 


1961 lava flow 
Settlement 


rae: as 
Om 
4 6760t 


>, &/ 
INACCESSIBLE A 


See | 
oa eee le 


% NIGHTINGALE 


Be voung volcanic centres 


5 


mil 
Terrestrial scale Siniles 


Contours in feet 


< PENGUIN | 


e Alvarez 


Meteorological 
Station 


Fig. 3 Map showing the Tristan da Cunha group and distances between the major islands. 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 337 


islands fall within an overall mean latitude and longitude of 37°S. 12°W., but the co-ordinates 
for the individual islands, taken either from the Admiralty Chart or from the South Atlantic 
Pilot, are as follows: 


Tristan da Cunha 372 0675S 1 le? OW: 
Inaccessible 37 160-3: 12 52" W. 
Nightingale Si ele. Ss 127 32" W. 
Middle a7 25 30" S. 2 28° 307 W. 
Stoltenhoff Bh Pigs 12° 29’ W. 
Gough 40° 19’ § ONS” OW: 


The islands are entirely volcanic in origin and are thought to have arisen from a series of 
isolated eruptions in mid-ocean. Geologically they are on the line of the mid-Atlantic ridge 
and are believed to have been formed during the Tertiary period. The geology of the group 
has been described in some detail by Dunne (1941). 

The climate of the islands is temperate and of the extreme oceanic type, the prevailing 
winds being westerly. The air temperature is around | 1-12°C throughout the year, and the 
mean annual rainfall, from such figures as are available, is 1656 mm for Tristan da Cunha (at 
the Settlement) and 3225 mm for Gough Island. It must be observed, however, that on 
Tristan da Cunha, where the Peak dominates the topography of the island, the rainfall 
consequently varies with position and altitude. Precipitation is greater on the western side of 
the island than on the east, and the rainfall on the Base and at the Peak is approximately 
one-and-a-half times and three times greater respectively than that around the Settlement 
(Wace & Holdgate, 1958: 595; Wace, 1961: 339). 

Sealers and whalers, both British and American, made the Tristan group the centre for 
their activities from the middle of the 18th century, with the result that many of the place 
names commemorate either their ships or the men who served in them. A comprehensive list 
of these place names is given for Tristan by Crawford (1941: 219-228). A good map in Booy 
(1957: 8-9) indicates them; for Gough Island the map in Holdgate (1958: 21 1-216) should be 
consulted. 


Tristan da Cunha 

Tristan da Cunha (Fig. 4) was uninhabited by man when it was first discovered in 1506 by a 
Portuguese admiral named Tristao da Cunha. This island, with less than 100 sq km of land 
surface, is the largest of the Tristan-Gough group. The main area which is a volcano, some 
13 km across and roughly octagonal in shape, is known as the Peak (2058 m). It rises to a 
gigantic cinder cone (Fig. 1) from about 1220 m up to the rim of the crater (now filled by a 
lake). The slopes below are deeply furrowed with ravines radiating out from the cone like the 
spokes of a wheel, and are covered with a dense vegetation of the tree fern Blechnum palmi- 
forme. These slopes level out at the foot to a plateau which the islanders refer to as the Base, 
although this is still at altitudes at 240-610 m above sea-level. This plain is variously 
covered with grass sward merging into a scrub of Phylica arborea covered with moss. The 
vegetation of the island has been well described by Wace & Holdgate (1958). The island has 
been continuously inhabited (Fig. 5) since 1810, apart from the 18-month break following 
the eruption of 1961, with cottages built upon part of the larva plain along the north-west 
side, in an area referred to as the Settlement. Vegetables, mainly potatoes, are cultivated 
nearby in the Potato Patches (Fig. 6) and cattle are grazed on the adjacent slopes. 


Inaccessible Island 

Inaccessible Island (Fig. 7), next in size to Tristan da Cunha and lying about 32 km to the 
south-west, was so named by Captain d’Etchevery of the French ship Etoile du Martin. 
Having in 1778 successfully set foot on both Tristan da Cunha and Nightingale Islands, 
d’Etchevery came to a third island but, failing to find a suitable landing place, he recorded 


338 E. W. GROVES 


BIG POINT a = 
ROOKERY POINT 


potato patches 


UNIVERSAL TRANVERSF MERCATOR GRID 


Fig. 4 Map of Tristan da Cunha. 


his failure to posterity by giving that island what he thought was the well-deserved name of 
Inaccessible. Landings have subsequently proved possible by using a beach along the north- 
east shore, although not without some degree of difficulty. The island is pear-shaped in 
general outline, being about 4.8 x 4 km, with a truncated volcanic cone rising to 560 m on 
the south-west side, and with a flattened plateau in the centre. Around the island precipitous 
cliffs drop sheer to the sea, but above, and rising to the peak, the plateau is variously covered 
with tall tussock grass (Spartina arundinacea), a scrub of the tree-fern (Blechnum palmiforme) 
and small areas of stunted Phylica arborea. A spectacular waterfall, several kilometres high, 
cascades from one of the perpendicular cliffs down to Salt Beach below (Fig. 8). Inaccessible 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 339 


pe 
2g Co bait be 
oa tig ee Re. ste 


Sas ee es ¢ = 4 - : . tt 
ae # “ : > ie “4 ; oo Le 
4 »-. at e! —I" =, moe » “ed eainte sg tl ES sg BR 


Fig. 5 David Hagen’s house on Tristan da Cunha, one of the oldest on the island. Note clumps of 
New Zealand flax, Phormium tenax, planted within the stone walls to act as wind-breaks. 
October 1966. Photograph: George Edwards. 


aes 


Fig. 6 The Potato Patches viewed from above on Hillpiece, Tristan da Cunha, and looking toward 
the western coastline of the island. October 1966. Photograph: George Edwards. 


340 E. W. GROVES 


INACCESSIBLE 


Sketch map—heights approximate only 
Arbitrary kilometre grid | 


BLENDEN HALL 


2km 


I miles 


6 
Fig. 7 Map of Inaccessible Island. 


Island is the only known locality in the world for the flightless rail (A¢t/antisia rogersi) or 
‘Island hen’ as named by the Tristan islanders (Green, 1973: 92), to distinguish it from a 
slightly larger bird formerly known on Tristan da Cunha itself called ‘Island cock’ (=Tristan 
moorhen) (Gallinula nesiotis nesiotis)—now unfortunately extinct. 


Nightingale Island 

Nightingale Island (Fig. 9) lying to the south-east of Inaccessible Island, and some 29 km 
south-south-west of Tristan da Cunha, is named after a British Naval Officer, Captain 
Gamaliel Nightingale, who visited the Tristan group in 1760. It is rectangular in shape, being 
about 1.6 km west to east and 1.2 km north to south. It has two peaks; the eastern one, more 
precipitous and conical, rises to 336 m, while the southern one reaches 292 m and has more 
gently sloping sides except on its southern flank. High cliffs bound its southern, eastern and 
western shores. The lower land and all but the steeper slopes are covered, as on Inaccessible 
Island, with tussock (Spartina arundinacea), the tree-fern Blechnum palmiforme and occa- 
sional thick patches of Phylica arborea. Moseley (1879: 108-109) described his difficulty in 
penetrating this vegetation and how he found that one large area, densely overgrown, had 
become a vast penguin rookery. It has been estimated that about two million pairs of the 
great shearwater (Puffinus gravis) nest in burrows on Nightingale Island (Rowan, 1952; 
Baird, 1965) and, together with those on Inaccessible and Gough Islands, form much the 
largest breeding colony for this species in the world. 


Middle Island 
Middle Island (Fig. 9) is situated less than 0.8 km north of Nightingale and is only about 
0.4 km square in size. It is low lying and flat-topped, the highest point being only 45 m. 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 34] 


ad ogee ee tae : = 
Fig. 8 Salt beach, Inaccessible Island, with cliffs down which tumbles a spectacular waterfall. 


These cliffs are mantled with tussock grass, Spartina arundinacea, on all but the steepest slopes. 
February 1968. Photograph: Nigel Wace. 


Most of the island is covered with a thick entanglement of tussock grass and is occupied by 
many penguin rookeries. 


Stoltenhoff Island 

Stoltenhoff Island (Fig. 9) was named after the brothers Stoltenhoff, crew members of a 
whaling ship, who had asked to be landed on Inaccessible Island thinking they would make 
their fortune there in fur sealing. After nearly two years of disappointment and spartan 
existence they were taken off by HMS Challenger when she called in October 1873. 
Stoltenhoff Island lies about 1.6 km north-west of Middle Island and is approximately 460 m 
long by 135 m wide. Its vertical cliffs, varying from 70 m to 80 m high, rise sheer from the 
sea. Only at one site, below the cliffs on the north-west side, is it possible to land, but even 
there this is hazardous because of the ocean swell. Like Middle Island, the centre is flat 
topped, rising to 100 m high, and is densely overgrown with tussock grass. 


Gough Island 

Gough Island (Fig. 10) is so called because it was sighted in 1731 by Captain Gough, master 
of the barque Richmond bound for China. It was no doubt, however, the same island as that 
reported by the Portuguese in the 16th century and named by them as Gongalo Alvarez after 
the captain of Vasco da Gama’s flagship (Holdgate, 1958: 75). Sealers lived ashore for 
varying periods of up to two years during the 19th century (Wace, 1961: 337). There has 
been a manned weather station on the island since 1955. 


342 E. W. GROVES 


STOLTENHOFF Is. 
p 


NIGHTINGALE ea ees 


1 : 
Sketch map. heights approximate only /2 mile 


4 
Fig. 9 Map of Nightingale Island with the subsidiary Stoltenhoff and Middle Islands to the north. 


The island lies about 320 km south-south-east of Tristan da Cunha and is approximately 
13 km long and 5-6 km wide. Steep cliffs skirt most of its coastline (Fig. 11), those at the 
north-east end being up to 370 m high. Rainfall, as previously mentioned, is high, and the 
waters of several falls drain down to the cliff edges and cascade precipitously into the sea. 
Landing is usually effected at the boulder beach at the entrance to a ravine known as the 
Glen (Fig. 12). The interior of the island is an undulating boggy plateau with an elevation of 
610 m from which several peaks rise, the highest being Edinburgh, 910 m. Vegetation covers 
all the island except for the sheerest cliffs and the tops of the peaks, and has been adequately 
described by Wace (1961). The island is the only locality for the flightless Gough moorhen 
called ‘Island cock’ Gallinula nesiotis comeri, a different subspecies to that formerly 
occurring on Tristan da Cunha. 


Visitors who have made vascular plant collections 


Chronological list 
The following list is a chronological list of persons, expeditions and ships known to have 
made vascular plant collections on the islands: 


343 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 


“‘purys] ysnoH jo de~ OF “Sd 


” 
“ 
ANIO¢d MINOS 


| = Z 
ANIOd 153 MINOS iw) WOLVIWIN ISWIASNYUL Wwsw3AInn 
| sopiw Zz t ° 


iw 
| QNOwnot 
O48 


t 


ave 
Tevas' VEL 


_NONVIS ea 
Ww315O1ONOR aH ( 


! 
INIO€ $ninovo 
1 


: NIVid oo 
Nes = wa SsOuively. Oved wise 
H | ave isan = 
i | ~ 7 WIOO1-SO¥H 1 am : 
\ $50 Pie “25>, { of 
i Lia 


Z3UWA TW OTWD NOD. 


al Nuvi 


= 1 Scere 
| Roh _ wad’ \ 
Sons ; 4 HOUNENIGI 
Fran) See 
| 7 


| a: Sail; 


| | eRe 
ce NG Dy | 


i Pee | 


MV3d INOMIG3dx2 


( t 

Bad ¢ ° 

oy. Va + a oe ALS 706, ic Be 
Wie a “4 - é ) 


’ Q $1aNO> 


WIA SIOTH ®& 


344 


E. W. GROVES 


Fig, nl East coast a Cons ae eee AG towards Reef Point, with Penguin Islet (115 m) 
lying 0.8 km off the point. The lower slopes of Pummel Crag in mid-foregrounnd are covered 
with the tree fern Blechnum palmiforme and thickets of Phylica arborea. May 1968. Photograph: 
Nigel Wace. 


3-7 January 1973 
1816-1817 

12 November 1852 
15 October 1873 
October 1873 
October 1874 

July 1892 
February 1904 
1908-1909 

20 May 1922 


1922-1925 

1926 

10 November 1934 
February—March 1937 


1937-1938 
February 1939 
1954-1956 


11-29 October 1955 

February & March 1962 
March 1968 

1971-1974 

November and December 1976 


Tristan da Cunha 


Aubert Du Petit-Thouars 

Carmichael 

Macgillivray & Milne (HMS Herald) 

Moseley (HMS Challenger) 

Crosbie 

Saunders 

Richardson 

Bonomi 

Keytel 

Wilkins (Shackleton-Rowett Expeditions in the SY Quest but see 
remarks on pp. 353 and 406 under Polystichum mohrioides.) 

Rey. and Mrs Rogers 

Glass 

Siggeson 

Dyer (HMS Carlisle) 

Christophersen & Mejland (Norwegian Scientific Expedition) 

MacCraken 

Stableford 

Wace (Gough Island Scientific Survey) 

Dickson (Royal Society Expedition) 

Wace 

Fleming 

Wace 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 


Fig. 12 Gough Island, looking up the Glen from Archway Rock to Hag’s Tooth and Mt Rowett. 
The steep slopes are covered with the ferns Blechnum palmiforme and Histiopteris incisa var. 
carmichaelina, and with thickets of Phylica arborea. May 1968. Photograph: Nigel Wace. 


346 


16 October 1873 
23 May 1922 
February 1938 
February 1957 
March 1962 
March 1968 


September 1973-—January 1974 


November 1976 


17 October 1873 
1922-1925 
February 1938 
1953-1954 


March 1968 
February 1972 
November 1976 


| February 1938 


8 February 


22 April 1904 

I June 1922 

8 June 1927 

27 February 1933 
March 1955 


28 November 1955-5 March 1956 


March 1956 
1956-1957 
February 1961 
May 1968 
January 1972 
November 1976 


Alphabetical list 


E. W. GROVES 


Inaccessible Island 
Moseley (HMS Challenger) 
Wilkins (Shackleton-Rowett Expedition in the SY Quest) 
Christophersen (Norwegian Scientific Expedition) 
Stableford 
Dickson (Royal Society’s Expedition) 
Wace 
Fleming 
Wace 


Nightingale Island 

Moseley (HMS Challenger) 

Rev. & Mrs Rogers 

Christophersen (Norwegian Scientific Expedition) 

Stableford (but see comments under alphabetical list of expeditions 
and visitors below regarding certain specimens labelled from 
Nightingale.) 

Wace 

Fleming 

Wace 


Middle Island 
Christophersen (Norwegian Scientific Expedition) 


Stoltenhoff Island 
Christophersen (Norwegian Scientific Expedition) 


Gough Island 
Brown (Scottish National Antarctic Expedition in the SY Scotia) 
Wilkins (Shackleton-Rowett Expedition in the SY Quest) 
‘Discovery’ (‘W.S.’) Expedition 
Christensen 
Swain 
Wace (Gough Island Scientific Survey) 
MacMillan 
van der Merwe 
McKinnon 
Wace 
Fleming 
Wace 


The following is an alphabetical list of personnel, expeditions and ships known to have made 
vascular plant collections on the islands, together with a note of the institutes where their 
collections are now deposited. Herbaria are denoted by their official abbreviations (see 
Holmgren & Keuken, 1974). The following collections were consulted during the prepara- 


tion of this paper: 


AAS = British Antarctic Survey (Botanical Section) Cambridge, England. 

BM = British Museum (Natural History), London, England. 

CGE = Botany School, The University, Cambridge, England. 

E = The Royal Botanic Garden, Edinburgh, Scotland. 

K = The Royal Botanic Gardens, Kew, Surrey, England. 

NBG = Compton Herbarium, National Botanic Gardens, Kirstenbosch, Cape Town, South 
Africa. 

O = Botanisk Museum, Oslo, Norway. 

Pe = Laboratoire de Phanérogamie, Museum National d’Histoire Naturelle, Paris, France. 


PRE = National Herbarium, Botanical Research Institute, Pretoria, South Africa. 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 347 


AUBERT Du PETIT-THOUARS, AUBERT 

A. Aubert Du Petit-Thouars (Fig. 13) was a traveller and botanist who visited Tristan da 
Cunha in 1793 when the vessel in which he was travelling to the Mascarenes made a five-day 
stay for watering. He was ashore on the island on the 3, 5 and 6 January, during which time 
he made the first known botanical investigation of the island and collected 100 species, 
including 25 flowering plants and 16 ferns (the remainder being non-vascular cryptogams). 
The first set of his material is at P with a few duplicates at BM. For the correct form of his 
name see Barnhart (1965: 1:88); in the citation of specimens it has been abbreviated to 
Thouars. Literature: Aubert Du Petit-Thouars (1808). 


BARKLY, HENRY 

Two ferns, Dryopteris aquilina and Elaphoglossum succisaefolium, gathered whilst HMS 
Challenger visited Tristan da Cunha in 1873, are associated with Sir H. Barkly although 
probably not collected by him. The specimens are in BM. 


BONOMI, P. 

Nothing is known of P. Bonomi beyond the fact that he visited Tristan da Cunha and made a 
collection of flowering plants and ferns there in February 1904. The first set of these 
specimens is in NBG (transferred from the old South African Museum), with a few dupli- 
cates at K. On some sheets in NBG his name appears as P. Benomi—perhaps an alternative 
spelling. 


BROWN, ROBERT NEAL RUDMOSE 

R. N. R. Brown was botanist on the Scottish National Antarctic Expedition to the Weddell 
Sea, 1902-04. On his return, the expedition’s ship SY Scotia, anchored for three days off 
Gough. Only on one day (22 April 1904) could a boat be got ashore because of continual 
high seas. Even then, because of the poor anchorage, the shore party had to keep within 
reach should sudden recall be necessary. Thus no visit to higher ground was possible. The 
first set of plants is at K, the second at E, and a few duplicates at BM. Literature: Brown 
(1905), Brown, Wright & Darbishire (1912). 


HMS Car.isLeE 
See under Dyer. 


CARMICHAEL, DUGALD 

Captain Dugald Carmichael of the 72nd Regiment landed on Tristan da Cunha on 28 
November 1816 with a British garrison sent to take possession of the island to prevent a 
possible attempt by the French to rescue Napoleon Bonaparte, imprisoned by the British on 
St Helena. Carmichael had been attached to the garrison at his own request in order to make 
a botanical survey of the island (Brander, 1940: 72). During the three months he was there he 
made a collection of vascular plants totalling 55 species (Hemsley, 1885: 138), some of which 
are at K and others at BM. Literature: Carmichael (1819), Hemsley (1885). 


HMS CHALLENGER 
See under Moseley. 


CHRISTENSEN, LARS 

Consul Lars Christensen, sponsor of several Norwegian Antarctic expeditions during the 
1920-1939 period, landed on Gough on 27 February 1933 from MS Thorshavn when the 
ship was returning from the Antarctic at the end of the 1932-33 season. The plants gathered 
during the brief visit ashore were sent to Dr Erling Christophersen in Oslo. He found that 
among the 22 species collected by Christensen, 12 were new to the island of which three were 
new to science. The first set is at O. Literature: Christophersen (1934). 


CHRISTOPHERSEN, ERLING and MEJLAND, YNGVAR 
The bulk of the vascular plants collected during the Norwegian Scientific Expedition to 
Tristan da Cunha, 1937-38, were gathered by Erling Christophersen (leader and botanist) 


348 E. W. GROVES 


(Fig. 16) although some were collected by Yngvar Mejland, botanical assistant and the 
expedition’s handyman. Most of the collecting was done on Tristan itself, where the primary 
work of the expedition was carried out, but time was also spent on Nightingale and Inacces- 
sible Islands during February 1938. No visit was possible to Gough Island, but a brief 
landing was made on both Middle and Stoltenhoff Islands and the few plants collected 
represent the first botanical specimens to be gathered from these subsidiary islands of 
Nightingale. The expedition departed from Tristan da Cunha on 29 March 1938. The first 
set of the vascular plants is at o, the second set at BM, the third at C and the fourth at K. 
Other smaller part sets are at P, PRE, BOL, GH and WELT. Although some duplicate sets of 
vascular plants of this expedition have both collectors’ names appearing on the accompany- 
ing labels (either printed or hand written), the original set (O), as well as most of the other 
distributed sets, bears simply the name of either Christophersen or Mejland on each label 
followed by the relevant number allocated to that collector in the field. Material from this 
expedition cited throughout this present paper follows that sequence, thus: 


Christophersen 1-100 1051-1109 
201 — 300 tii 11338 
351 -—600 1181-1286 
628 -— 631 1312 - 1335 
642 1377 
644 - 810 1652b 
825 2000 — 2236 
896 (Nightingale Is.) 
945 2300 — 3000 
955 (Inaccessible Is.) 
Mejland 101 — 200 1041 — 1050 
301 — 350 1110 
601 — 627 1134-1179 
632 - 641 1287 — 1309 
643 1356 — 1576 
811 —824 1591 — 1650 
1011-1014 1654-1738 
1017-1038 1745 — 1820 


Literature: Narrative: Christopherson (1940), Crawford (1941); Botanical: Christensen, C. 
(1940), Christophersen (1934, 1937, 1940b, 1944, 1968), Hooper (1968). 


CROSBIE, A. 

Dr A. Crosbie was staff surgeon in the Royal Navy on board HMS Challenger during that 
vessel’s oceanographic expedition of 1872-1876. He collected a few ferns (including 
Blechnum australe and Huperzia insularis) on Tristan da Cunha when a landing was made 
there in October 1873. These subsequently became the property of William Evans and were 
presented by his son to the herbarium at E in 1933. 


DICKSON, JAMES HOLMS 

J. H. Dickson was botanist with the Royal Society’s expedition to Tristan da Cunha in 
1962, to assess the damage caused to both the flora and fauna by the eruption of the island’s 
volcano in 1961. Vascular plants, together with bryophytes and lichens, were collected 
between 2 February and 17 March 1962 mostly on Tristan da Cunha, but a visit was also 
made to Nightingale (1 & 2 March) and Inaccessible (4 March). The first set of vascular 
plants is at BM with a few duplicates at K, AAS and CGE. Literature: Dickson (1965a), 
Dickson (19655), Wace and Dickson (1965). 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 349 


pa vA Se A 
Senne Sut irciilleire: 
fee $rance ; va 


Fig. 13 Aubert Aubert Du Petit-Thouars. Fig. 14 John Macgillivray. 


Fig. 15 Henry Nottidge Moseley. Fig. 16 Erling Christophersen. 


350 E. W. GROVES 


DiscOVERY INVESTIGATIONS 

During the return from the 1926-27 Discovery investigations, the RRS William Scoresby, 
which had been on whale-marking and oceanographical surveys with RRS Discovery IT off 
South Georgia and the Falklands, called in at Gough Island on 8 June 1927. Sir Alister 
Hardy (1967: 433-437) marine zoologist, writing of this cruise, describes the landing and 
subsequent climb up to about 370 m along a central ridge leading up to Mt Howett. His two 
companions, who collected both zoological and botanical material, are identified (Hardy, 
1967: 434) as Dilwyn John (zoologist) and Rolfe Gunther (zoologist). The latter, in his 
account of the whole expedition (Gunther, 1928: 41), also mentions this short visit to Gough 
Island. The plant material (allocated Station No. 123) was brought back in formalin, as was 
usual with zoological specimens, and was presented to the BM, where it was subsequently 
dried and placed in the herbarium. Literature: Hardy (1967), Gunther (1928). 


Du PETIT-THOUARS 
See under Aubert Du Petit-Thouars. 


DYER, ROBERT ALLEN 

R. A. Dyer (Fig. 17) was seconded from the South African Government’s Division of Botany 
to join the HMS Carlisle expedition to Tristan da Cunha in 1937. The ship had been sent by 
the British Admiralty to take provisions and mail to the island, to investigate the general 
conditions and health of the inhabitants, and to embark the resident minister, the Rev. 
H. Wilde, for leave. It arrived from Cape Town on 28 February 1937 and stayed three days. 
Plant collections were made ashore by Dyer on two days; on the | March near the beach 
landing stage and margin of the Settlement Plateau and on the 2 March (with the help of an 
islander, Arthur Rogers) up the slopes of the Peak to 760 m. Dyer (1939) records the species 
he collected, and gives a list of some pasture grasses and species of trees and shrubs brought 
from South Africa that were intended to be established on the island. The first set of Dyer’s 
Tristan plants is at PRE, the second at NBG; smaller part sets are at BM, K and O. 
Literature: Dyer (1939). 


FLEMING, J. I. H. 

J. I. H. Fleming served as Government Administrator on Tristan da Cunha from 1971 to 
1974 and made a collection of about 100 numbers of flowering plants and ferns, not only 
from Tristan da Cunha, but also from Gough Island (January 1972), Nightingale Island 
(February 1972) and Inaccessible Island (September 1973 to January 1974). His specimens 
are at E. 


GLASS, DONALD 

D. Glass was a Tristan islander who collected a few flowering plants and ferns on Tristan da 
Cunha in 1926. These were subsequently forwarded to Mrs Meade-Waldo of Edenbridge, 
Kent, who presented them to the BM in 1927. 


GOUGH ISLAND SCIENTIFIC SURVEY, 1955-56 
See under Wace. 


HMS Heratp 
See under Macgillivray and Milne. 


KEYTEbabaC: 

P. C. Keytel visited Tristan da Cunha on two occasions: first in 1907, then for a more 
prolonged stay in March 1908 to negotiate a trade in sheep and dried fish between the 
islanders and the South African market (Munch, 1971: 70-80). His second visit lasted a year 
(1908-09), during which time he made a botanical collection of 57 species (43 flowering 
plants and 27 ferns) (Phillips, 1913). The first set of his material is at NBG (transferred, like 
those of P. Bonomi (q.v.), from the old South African Museum); and a few duplicates are at 
K. Literature: Phillips (1913). 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 351 


McCRAKEN, M. L. 
Nothing is known of McCraken except that he gathered a specimen of Chenopodium 
tomentosum on Tristan da Cunha in February 1939 which is now at BM. 


MACGILLIVRAY, JOHN and MILNE, WILLIAM GRANT 

The two collectors John Macgillivray (Fig. 14) and William G. Milne were naturalist and 
assistant naturalist respectively on board HMS Herald (under the command of Captain H. 
M. Denham) during that ship’s second surveying voyage in 1852-61 to the Pacific and Indo- 
Pacific waters. On the outward voyage the ship called at Tristan da Cunha in November 
1852 and it seems that both naturalists collected at the same localities while ashore. 
According to Hemsley (1885: 139) about 40 species of vascular plants, together with a few 
mosses and hepatics, were collected. Macgillivray’s specimens are mainly at BM and a few 
are at K; Milne’s specimens are at K with possibly some at CGE. Those at BM and K include 
both a numbered and an unnumbered series. It should be noted that Macgillivray (and his 
father) wrote his name in this way, and not as ‘MacGillivray’ (Barnhart, 1965: 2:424). 
Literature: Botanical specimens are cited throughout Hemsley (1885). 


McKINNON, R. S. 
R. S. McKinnon, whilst serving as a naval officer on the Royal Naval Survey vessel HMS 
Owen, collected a few numbers of ferns on Gough Island in February 1961. These are now at 


MACMILLAN, R. A. B. 
MacMillan visited Gough Island in May 1956 and gathered a few numbers of flowering 
plants and ferns which are now at PRE. 


MEJLAND, YNGVAR 
See under Christophersen and Mejland. 


MERWE, J. J. VAN DER 

J. J. van der Merwe was meteorologist to the Gough Island Scientific Survey expedition 
1955-56. He then remained on that island for a further two years in charge of the meteoro- 
logical station, set up by the South African Weather Bureau, after the expedition had left. 
During 1956-58 van der Merwe made a small collection of plants, the first set of which is 
at PRE, with a few duplicates at K and BM. Literature: A few specimens are cited in Wace & 
Dickson (1965: Appendix A). 


MILNE, WILLIAM GRANT 
See under Macgillivray and Milne. 


MosELey, HENRY NOTTIDGE 

H. N. Moseley (Fig. 15) was appointed naturalist to HMS Challenger during its world ocean- 
ographic voyage of 1872-76. On the outward voyage the ship called at the Tristan group of 
islands. It arrived at Tristan da Cunha on 15 October 1873 and immediately Moseley was 
able to embark he began botanizing. He was ashore for six hours before being recalled to the 
ship because of a heavy squall. The ship visited Inaccessible Island the next day (16 October) 
and Nightingale Island on 17 October, before sailing for Cape Town on 18 October 1873. 
Hemsley (1885: 139) states that Moseley collected in all 45 species of vascular plants from 
the three islands he visited (38 species on Tristan da Cunha, 21 species on Inaccessible 
Island, and 10 species on Nightingale Island). Of the vascular plants, the first set seems to be 
at K and the second at BM, with a few duplicates at E. A few Moseley specimens from these 
islands are also at CGE (see Gilmour and Tutin, 1933: 24). Literature: Narratives: Spry 
(1877), Moseley (1879 and subsequent editions), Linklater (1972); botanical: Moseley 
(1874), Hemsley (1885). 


NORWEGIAN SCIENTIFIC EXPEDITION TO TRISTAN DA CUNHA, 1937-38 
See under Christophersen and Mejland. 


359 E. W. GROVES 


RICHARDSON, J. ; 
Nothing is known of Richardson except that he gathered the fern Lycopodium diaphanum 
on Tristan da Cunha in July 1892. The specimen is now in O. 


Rocers, Henry M. and RoGERS, Ross A. 

Rey. H. M. Rogers was missionary-priest on Tristan da Cunha from | April 1922 until 4 
February 1925. While there he made, together with his wife, a small collection of plants 
consisting of 20 numbers (12 flowering plants, 7 ferns and | moss), which were presented to 
BM, with a few duplicates to K. The identification of this collection appeared as part (pp. 
212-213) of Appendix II to the book later written by Mrs Rogers on their experiences during 
three years stay on the island. The Rev. Rogers died on 14 May 1926. Literature: Rogers 
(1926). 


ROYAL SOCIETY EXPEDITION TO TRISTAN DA CUNHA, 1962 
See under Dickson. 


SAUNDERS, W. W. 
Nothing is known of W. W. Saunders except that he gathered a specimen of Rumex angio- 
carpus on Tristan da Cunha in October 1874 which is now at K. 


SCOTTISH NATIONAL ANTARCTIC EXPEDITION, 1902-04 
See under Brown. 


SHACKLETON-ROWETT ANTARCTIC EXPEDITION, 1921-22 
See under Wilkins. 


SIGGESON, EINAR 

On the voyage to the southern Antarctic whaling grounds during the 1934-35 season, the 
Norwegian whaler Thorshammer made a stop at Tristan da Cunha on 10 November 1934 to 
put ashore Mr Einar Siggeson, a whaling captain, who, at the the request of Consul Lars 
Christensen, had been asked to collect plants and animals. The plants thus gathered were 
forwarded to the Botanical Museum at Oslo in the spring of the following year. When 
examined by Dr Erling Christophersen, the collection was found to include 36 species 
including 17 flowering plants and 5 ferns, one of which was new to the Tristan flora. 
Literature: Christophersen (1937: 8-13). 


STABLEFORD, H. GERALD 

H. G. Stableford served as agricultural officer on Tristan da Cunha from 1953-57 (Munch, 
1971: 246). While there he visited other islands in the group and on his return to England 
brought back flowering plants and ferns that he had gathered on Tristan da Cunha 
(1954-56), Inaccessible Island (February 1957) and Nightingale Island (1954). A few, 
although not all, of the herbarium specimens at K labelled to have come from Nightingale 
were undoubtedly collected on Tristan. Two sets of plants were gathered, although the 
material in each does not seem to be mutually exclusive. One set was deposited at BM, and 
the other at K. The BM set was found to have several duplicated series of numberings, and so 
after receipt the whole collection was in consequence renumbered as a single consecutive 
series 1n order to avoid future confusion. The BM now has numbers running from 1-132. 
The set of plants at K consists of a numbered series (several numbers being repeated) as well 
as unnumbered series, presumably in much the same condition as it was received. The 
present situation is unfortunate because material of the same gathering in both herbaria often 
does not bear the same number. Literature: Some specimens are cited in Wace & Dickson 
(1965: Appendix A). 

SWAIN, ADAM 

An ieee who in March 1955 collected a few ferns on Gough Island, now in the herbarium 
at K. 

THOUARS 

see under Aubert Du Petit-Thouars. 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 353 


Fig. 17 Robert Allen Dyer. Fig. 18 Nigel M. Wace. 


WACE, NIGEL M. 

N. M. Wace (Fig. 18) was botanist to the Gough Island Scientific Survey of 1955-56. The 
expedition first landed on Tristan da Cunha, where they stayed from 1 October to 13 
November 1955 and from whence they sailed in the Tristania (a 600-ton fishing vessel 
belonging to the Tristan da Cunha Development Company) to Gough Island. The 
expedition remained on Gough Island for six months from 14 November 1955 until they 
were taken off the South African naval frigate Transvaal on 13 May 1956, and returned then 
to Tristan da Cunha briefly before sailing for Cape Town. An account of the work of the 
entire expedition has been published by Holdgate (1958). Regarding the vascular plants 
collected during the expedition, two series were gathered: 94 numbers from Tristan da 
Cunha (each prefixed with the letter T) and 163 numbers from Gough Island (without 
prefix). The first set of all numbers is at BM. Wace visited the Tristan group again from 
March to May 1968 during which period he was able to make further collections on Tristan 
da Cunha (including many species colonizing the volcanic debris resulting from the eruption 
of 1961), and on Inaccessble, Nightingale and Gough Islands. A further expedition to these 
four major islands was made by him from November to December 1976. At the time of 
writing (1979) duplicates of only the grasses, sedges and rushes have so far been distributed 
from the 1968 and 1976 expedition, specimens of which have been lodged at Kew. 
Literature: Narrative: Holdgate (1958); Botanical: Wace & Holdgate (1958), Wace (1961). 


WILKINS, GEORGE HUBERT 
G. H. Wilkins was appointed naturalist to the Shackleton-Rowett Antarctic Expedition of 


1921-22. On returning from the Antarctic, the expedition’s vessel, the SY Quest, visited 
several islands of the Tristan-Gough group: Tristan da Cunha (20 May 1922), Nightingale 
Island (21 May), Inaccessible Island (23 May) and Gough Island (28 May to 2 June), before 
sailing for Cape Town on 3 June 1922. No vascular plants were collected on Tristan da 
Cunha (except perhaps for the fern Polystichum mohrioides, see p. 406) or on Nightingale 


354 E. W. GROVES 


Island, but a small collection was made on Inaccessible Island as well as on Gough Island. 
Wilkins was able to go ashore on the latter for five out of the six days they were there, 
although landing was difficult in deteriorating weather conditions (Wild, 1923: 265-286). 
Wilkins (1925: 68-70) lists those species (including 13 flowering plants and 9 ferns) he 
collected on Gough Island, but not all of these have since been traced at BM, to which the 
whole of the Shackleton-Rowett expedition’s plant collections were presented. Literature: 
Narrative: Wild (1923); Botanical: Wilkins (1925). 


RRS WiILL14mM SCORESBY 
See under Discovery Investigations. 


Systematic list 


SPERMATOPHYTA 


The families are arranged by a modified Bentham & Hooker system. The species appear 
alphabetically within each family. 


RANUNCULACEAE 


[Ranunculus acris L. 

This species is given as occurring on Tristan da Cunha in Wace & Dickson (1965; 334 and in Appendix 
A) and said to be based on an unpublished collection in BM. Unfortunately no material in that 
herbarium has been traced to substantiate this record.] 


Ranunculus carolii Christoph. 

Endemic. This species was described by Christophersen (1950: 186) from a specimen found 
by Mejland on Tristan during the Norwegian Scientific Expedition 1937-38. However, a 
specimen collected on that island by Carmichael in 1817, since identified as this species, 1s 
the earliest record for the group. It has also been found on Gough by Wace in 1955. The 
species is regarded by Christophersen (1968: 8) as being endemic and closely related to R. 
biternatus Sm. (R. crassipes J. D. Hook.) 

Tristan: Carmichael s.n. (BM); Christophersen 60 (O), 550d (O) 1777 (O); Mejland 1369 
(O-holotype, BM), 1558 (O); Wace T.71 (BM). Nightingale: Stableford s.n. (K) [Likely to be in error for 
Tristan]. Gough: Wace 26 (BM), 28 (BM), 51 (BM), 78 (BM), 144 (BM). 

Ranunculus repens L. 
Introduced. First noted from Tristan da Cunha in 1938 as a weed near a garden and also at 
the Potato Patches. 

Tristan: Christophersen 1788 (O, BM); Mejland 1288 (O, BM). 


CRUCIFERAE 


Brassica juncea (L.) Czernj. 
Introduced. Found on Tristan da Cunha in pasture at Anchorstock Bay in 1938. 
Tristan: Mejland 1411 (O, BM). 


Brassica rapa L. 
Introduced. Found as weed in gardens at the Settlement on Tristan da Cunha in 1938 and 
around ruined cottage at Saltbeach, Inaccessible Island, in 1962. 

Tristan: Mejland 340 (O, BM) [under Brassica campestris L. in Christophersen, 1968: 9]. 
Inaccessible: Dickson 127 (BM), 128 (BM). 


Cardamine glacialis (G. Forster) DC. 

C. propinqua Carmich. 
Native. First discovered on Tristan da Cunha by Carmichael in 1817 and described by him 
as C. propinqua (Carmichael, 1819: 507). This was later found to be the same as C. glacialis 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 355 


originally described by G. Forster as Sisymbrium glaciale in 1789 from material collected in 
Tierra del Fuego 

Tristan: Carmichael s.n. (BM-holotype of C. propinqua); Christophersen 517 (O), 534 (O), 
1275 (O); Dickson 109 (BM). Inaccessible: Christophersen 2348 (O); 2500 (O). Gough: Lars 
Christensen s.n. (O); van der Merwe 26 (K); Wace 25 (BM), 25a (BM), 145 (BM). 


Coronopus didymus (L.) Sm. 
Introduced. First recorded from Tristan da Cunha by Stableford near the Settlement in 1954. 
Tristan: Stableford s.n. (K), 7 (BM); Dickson 19 (BM). 


Nasturtium officinale R.Br. 
Introduced. Discovered on Tristan da Cunha by Dickson in 1962 by the Big Watron 
Waterfall* near the Settlement. 

Tristan: Dickson 144 (BM)t. 


Raphanus sativus L. 
Introduced. Aubert Du Petit-Thouars (1808: 46) reported this species from Tristan da 
Cunha, growing in a small area which was at one time under cultivation. Herbarium material 
from Inaccessible Island, which was collected at Saltbeach by Christophersen in 1937, is no 
doubt also a relic of former agrarian activities. 

Tristan: Thouars s.n. (P). Inaccessible: Christophersen 2622 (O). 


CARYOPHYLLACEAE 


Cerastium fontanum subsp. triviale (Link) Jalas 

C. triviale Link; C. caespitosum subsp. triviale (Link) Hiit.; C. caespitosum Gilib.; C. holosteoides Fr. 
Introduced. Has been found in small quantity in grassland turf of the Settlement Plain on 
Tristan da Cunha, and also in several localities at Blenden Hall on Inaccessible Island. It was 
found on Gough Island for the first time in 1972. 

Tristan: Bonomi 16 (NBG); s.n. [herb. Mus. Austro-Afric. 1145] (NBG); Christophersen 13 
(O, BM), 44 (O, BM); Dickson 76 (BM) 94 (BM); Dyer 3532 (PRE, K); Keytel 1792 (NBG, K); 
Macgillivray 352 (K); Mejland 175 (O), 611 (O); Moseley s.n. (BM), s.n. (K); Siggeson 40 (O); 
Stableford 17 (BM), 105 (BM), s.n. (K); Wace T.72 (BM), T.79 (BM). Inaccessible: Christophersen 
2303 (O, BM), 2381 (O); Moseley s.n. (BM, K, E). Gough: Fleming 2 (E). 


Polycarpon tetraphyllum (L.) L. 

Introduced. First recorded from Tristan da Cunha by Macgillivray during the visit of HMS 

Herald to the island in November 1852 and since found in pastures and on open ground. 
Tristan: Christophersen 276 (O); Dickson 35 (BM); Macgillivray 351 (K); Mejland 120 (O, 

BM), 126b (O); Moseley s.n. (K, E); Stableford s.n. (K). 


Scleranthus annuus L. 
Introduced. Recorded on Tristan da Cunha by Stableford in 1954 in the grounds of the 
Settlement mission. 

Tristan: Stableford 5 (BM). 


Silene alba (Mill.) Krause 
Introduced. Found on Tristan da Cunha during 1955-56, occurring as a weed at the 
Settlement. 

Tristan: Stableford 119 (BM). 


Silene gallica L. 
Introduced. Discovered on Tristan da Cunha during the summer of 1954 in the Mission 
garden at the Settlement, where it occurred as a weed. 
Tristan: Stableford 2 (BM). 
* Wace (in litt.) states that this is not called ‘Big Watron Waterfall’ now as the 1961 eruption altered the 


topography of the area. : ; ; 
No specimen has been traced which substantiates Stableford’s record of this species from Tristan da Cunha as 


listed in Wace & Dickson (1965, Appendix A). 


356 E. W. GROVES 


Spergula arvensis L. 
Introduced. Discovered on Tristan da Cunha during 1955-56 growing as a weed at the Settle- 
ment; probably introduced with imported seed. 

Tristan: Stableford 110 (BM). 


Stellaria media (L.) Vill. 
Introduced. Originally collected on Tristan da Cunha by Christophersen in 1937 as a weed 
in the Potato Patches and then on Gough Island by Wace in 1955, growing between boulders 
and under Rumex obtusifolius above the beach. 

Tristan: Christophersen 291 (O, BM), 295 (O, BM); Mejland 171 (O).* Gough: van der Merwe 59 
(PRE); Wace 4 (BM). 


MALVACEAE 


Malva parviflora L. 
Introduced. First recorded in pastures on Tristan da Cunha by Macgillivray in 1852. It has 
since been found in a meadow and a garden near the Settlement, presumably in the latter site 
as a weed. 

Tristan: Dickson 45b (BM); Macgillivray sm. (K); Mejland 101 (O, BM), 1293 (O); 
Siggesson 41 (O); Stableford 14 (BM). 
[Malva sylvestris L. 
This species is recorded from Tristan by Wace & Dickson (1965, Appendix A) who cite Christophersen 
(1937). Christophersen, however, merely quotes Hemsley’s reference to this species (Hemsley, 1885: 
145). The latter author listed introduced species known at that date from the Tristan group based on 
material in the collections of Aubert Du Petit-Thouars, Carmichael, Macgillivray, Milne and Moseley. 
The material of all these collectors is still extant but only one of them collected a Ma/lva. This specimen, 
which was gathered by Macgillivray and is in the Kew herbarium, is not Malva sylvestris but M. 
parviflora (q.v.).] 


GERANIACEAE 


Geranium dissectum L. 
Introduced. First recorded on Tristan da Cunha by the Norwegian Scientific Expedition of 
1937-38 as a weed occurring outside the Settlement gardens and in pasture at the Potato 
Patches. 

Tristan: Christophersen 162 (O, BM) 293 (O), 1290, (O, BM). 


Pelargonium grossularioides (L.) Aiton 

P. acugnaticum Thouars; P. australe var. acugnaticum (Thouars) Hemsley 

Native. First recorded by Aubert Du Petit-Thouars on Tristan da Cunha in 1793 and 
described by him as P. acugnaticum. It was subsequently reduced to a variety of P. australe 
Willd. by Hemsley (1885: 147) and identified with P. grossularioides by Kunth (1912: 410). 
Christophersen (1968: 13), however, still considers, as did Aubert Du Petit-Thouars, that it 
is taxonomically distinct; until further critical work it is perhaps best to retain it under P. 
grossularioides. 

Tristan: Bonomi 13 (NBG [herb. Mus. Austro-Afric. 1144], K,E); Carmichael s.n. (BM, 
K); Christophersen 244 (O), 277 (O. PRE); Dickson 15 (BM); Dyer 3547 (PRE, NBG, BM, K, 
O); Glass s.n. (BM); Keytel 1793 (NGB, K, BM), s.n. (PRE [herb. Marloth 4720]); Mejland 338 (O, 
BM); Moseley s.n. (K); Rogers 7 (BM), s.n. (K); Stableford 20 (BM), s.n. (K); Thouars s.n. (P-holotype 
of P. acugnaticum, BM). Inaccessible: Christophersen 2572 (O), 2547 (O, BM); Moseley s.n. (BM, K); 
Stableford 132 (BM). 


* No specimen has been traced to substantiate the Dickson record of the species from Tristan da Cunha as listed in 
Wace & Dickson (1965, Appendix A). 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 357 
OXALIDACEAE 


Oxalis corniculata L. 

Introduced. Occasional near houses or as a garden weed on Tristan da Cunha, and on 
Inaccessible Island. Christophersen (1968: 13) states that this cosmopolitan species was first 
collected on Tristan da Cunha by Macgillivray in 1852, quoting Hemsley (1885: 147). 
Hemsley (loc. cit.) cites also a Moseley specimen from the Tristan da Cunha and one from 
Inaccessible Island. Voucher specimens for these records are at K and BM. 

Tristan: Christophersen 124b (O, BM); Dyer 3535 (PRE); Green (for Rogers) 12 (BM); 
Keytel 1795 (NBG), s.n. (NBG [herb. Mus. Austro-Afric. 14609]); Macgillivray 350 (K); Mejland 106 
(O); Milne s.n. (K); Moseley s.n. (K, BM); Siggesson 44 (O); Stableford 8 (K), 23 (BM). Inaccessible: 
Christophersen 2619 (O); Dickson 129 (BM); Moseley s.n. (K). 


Oxalis purpurea L. TRISTAN SUNFLOWER 
O. variabilis Jacq. 
Introduced. First recorded from Tristan da Cunha by Keytel during 1908-09. This species 
occurs occasionally near houses at the Settlement and in the pasture near Waterun. Wace in 
1955 notes that it was apparently spreading only vegetatively. 
Tristan: Keyte/ 1794 (NGB, K, BM); Stableford 11 (BM), s.n. (K); Wace T.85 (BM). 


RHAMNACEAE 


Phylica arborea Thouars ISLAND TREE 

P. nitida auct. 

Native. First collected on Tristan da Cunha in January 1793 by Aubert Du Petit-Thouars, 
who based his type description on the material he gathered there (Aubert Du Petit-Thouars, 
1811: 45). Along with the ferns Histiopteris incisa var. carmichaeliana and Blechnum 
palmiforme, this semi-procumbent evergreen tree forms dense communities at lower alti- 
tudes on most of the islands. Most of the larger growth near the coastal strips on Tristan da 
Cunha has, in the past, been cut down, as it provided (and still does) a readily available 
source of fuel (Fig. 19). 

Tristan: Bonomi 20 (NBG [herb. Mus. Austro-Afric. 1143], K); Carmichael s.n. (BM), s.n. 
(K); Christophersen 426 (O, BM), 1133 (O, BM), 1816 (O); Dickson 11 (BM), 79 (BM); Dyer 3552 
(PRE, NBG, BM, K, O); Fleming 43 (E), 44 (E); Keytel 1796 (NBG, K, BM), s.n. (PRE [ex Herb 
Marloth]); Macgillivray 357 (K); Mejland 613 (O), 1137 (O), 1412 (O, BM), 1816 (O); Milne 7 (K); 
Moseley s.n. (K, BM, E); Rogers s.n. (fasciated) (K); Stableford 55 (BM); Thouars s.n. (P-holotype); 
Wace T 20 (BM). Inaccessible: Christophersen 2309 (O, BM), 2441 (O), 2442 (O), 2547 (O, BM); 
Moseley s.n. (K, BM, E. CGE). Nightingale: Christophersen 2062 (O, BM), 2105 (O, BM); Fleming 58 
(E); Moseley s.n. (BM, K); Rogers s.n. (K); Stableford 6 (K). Stoltenhoff: Christophersen 2143 (O), 2147 
(O, BM). Gough: Brown s.n. (K, E); Christensen s.n. (O); Discovery ‘W.S’ Exped. s.n. (BM); MacMillan 
1 (PRE); van der Merwe 37 (PRE), 50 (PRE, K); Wace 63 (BM), 103 (BM), Wilkins 38 (BM). 


LEGUMINOSAE 
Medicago polymorpha L. 
M. denticulata Willd. 
Introduced. Found growing on manure in the bull pen near the Mission gardens on Tristan 
during the summer of 1954. Stableford, who discovered it, considers it was probably 
introduced with imported dried fodder from South Africa. 
Tristan: Stableford 4 (BM, K), 4a (K). 


Medicago sativa L. 
Introduced. Collected on Tristan by Stableford in the summer of 1954 at Sandy Point, and 
believed by him to have been introduced to the island in 1953 with imported fodder. 

Tristan: Stableford 6 (K), 39 (BM), s.n. (K). 


Sophora microphylla Ait. 
Edwardsia macnabiana Graham; Sophora tetraptera var. microphylla J. D. Hook.; Sophora 
tetraptera forma goughensis E. G. Baker & Wilk.; S. macnabiana auct. 


358 E. W. GROVES 


Fig. 19 Gilbert Lavarello transporting on a donkey wood of the island tree, Phylica arborea, used 
for fuel. Tristan da Cunha. October 1966. Photograph: George Edwards. 


- 
ee 


Native. Recorded only on Gough Island. Brown, Wright & Darbishire (1912: 33) mention 
that a sealer, George Comer, kept a diary during his six-month stay on the island with his 
companions from August 1888 to January 1889. In it he noted ‘two kinds of trees’ he found 
growing in the interior (Verrill, 1895). Brown, Wright & Darbishire (loc. cit.) point out that 
one of these trees was undoubtedly Phylica nitida (= arborea), which would have certainly 
been familiar to Comer and the other sealers, and also to the islanders of Tristan da Cunha, 
from its presence on the latter island as a main source of fuel. The identity of the second tree 
is most likely to have been Sophora microphylla. Herbarium material of this Sophora from 
Gough Island was first collected by Wilkins (naturalist on the Shackleton-Rowett 
Antarctic Expedition) in June 1922, and his material became the type of the form goughensis 
which he described with Baker (Wilkins, 1925: 68). Sykes & Godley, after a study of the 
long-distance ocean dispersal of the southern hemisphere Sophora species, along with a 
comparison of plants raised from seed in their experimental garden at Christchurch, New 
Zealand, concluded that the Gough Island species was not S. tetraptera, S. prostrata or S. 
macnabiana (as has been designated at various times) but only a race of S. microphylla. 
Furthermore the Gough Island plants (along with those from southern Chile and Chatham 
Islands), although lacking the juvenile form of some New Zealand races of S. microphylla, 
‘differ no more than the New Zealand races do one from another’ and all have the same 
chromosome number 2n = 18 (Sykes & Godley, 1968: 495). Markham & Godley (1972: 639) 
consider that source of seed for the Gough Island population originated from Chile. 

Gough: van der Merwe 27a (BM) 38 (K); Wace 27 (M); Wilkins 92 (BM -syntype of forma 
goughensis), 93 (BM -syntype of forma goughensis). 
Trifolium dubium Sibth. 
Introduced. First recorded on Tristan da Cunha in 1937 in pasture meadow, at the 
Settlement and at the Potato Patches. The plant has been found several times on the island 
since that date. 

Tristan: Christophersen 273 (O, BM), 286 (O, BM): Dickson s.n. (BM); Mejland 136b (O), 163 (O, 
BM); Stableford 10 (K), 22 (BM), 104 (BM). 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 359 


Trifolium micranthum Viv. 
Introduced. Recorded on Tristan da Cunha by Wace in 1955 as abundant in close-grazed 
pastures around the Settlement. 
Tristan: Wace T.92 (BM). 
Trifolium pratense L. 
Introduced. First noted on Tristan by Dyer in 1937 growing in pasture (Dyer, 1939: 604) 
although the first substantiative herbarium specimen was not collected until 1953-54 by 
Stableford, who considered that its presence on the island was due to imported seed. 
Tristan: Stableford 3 (K). 
Trifolium repens L. 
Introduced. The earliest collection was made by Bonomi on Tristan da Cunha in 1904. 
Tristan: Bonomi s.n. (NBG [herb. Mus. Austro-Afric. 1142]; Christophersen 275 (O), 1453 (O); 
Keytel 1812 (NGB, K, BM); Mejland 166 (O, BM); Stableford 5 (K). 
Trifolium subterraneum L. 
Introduced. This plant is noted in a recent paper by Wace & Holdgate (1976: 45) as being a 
species introduced with others ‘as an impurity amongst grass seed for the playing field’ at the 
Settlement on Tristan da Cunha. It seems that in 1968 it was confined to that locality, 
although as those authors have pointed out ‘it may in future play an important part in the 
islander’s economy by increasing the productivity of the pastures’ (Wace & Holdgate, loc. 
Cit.). 
Ulex europeaeus L. 
Introduced. Originally noted from Tristan da Cunha by Mrs F. M. Barrow, who states that 
there were ‘a few bushes only near the houses’ (Barrow, 1910: 277). The only herbarium 
specimen appears to be that of Mejland, 1937. 
Tristan: Mejland 141 (O, BM). 


ROSACEAE 


Acaena sarmentosa (Thouars) Carmich. DOG CATCHER 

A sanguisorbae auct., non Vahl: Ancistrum sarmentosum Thouars. 

Native. This procumbent subshrub had been recorded from the four major islands, where it 
usually occurs amongst grass in open areas. Bitter (1911: 277-278) distinguished three 
varieties, the differences being mainly based on size and shape of the leaves and leaflets: a 
large-leaved variety, var /ongiuscula Bitter from Inaccessible Island; a small-leaved variety, 
var. /usciniae Bitter from Nightingale Island; and an intermediate variety, var tristanensis 
Bitter* from Tristan da Cunha, having leaves midway in size between the other two. These 
differences are not, however, accepted by Christophersen (1968: 10), as the material he 
collected on Nightingale and Inaccessible Islands, as well as that of other collectors cited 
below (most of which he examined), does not agree satisfactorily with Bitter’s varietal 
segregates. 

Tristan: Carmichael s.n. (BM), s.n. (K); Dickson 117 (BM); Dyer 3549 (PRE, NBG, BM, 
K, O); Fleming 82 (E); Glass s.n. (BM); Keytel 1797 (NBG, K, BM): s.n. (PRE [herb. Marloth 
4719]); Macgillivray 355 (K), s.n. (BM); Mejland 345 (O, BM, PRE); 641 (O); 815 (O); Milne 
s.n. (K); Moseley s.n. (K. BM); Rogers 6 (BM); Stableford 57 (BM); Thouars s.n. (P -holotype 
‘of Ancistrum sarmentosum, BM); Wace T.68 (BM). Imaccessible: Christophersen 2552 (O, 
BM); Moseley s.n. (K, BM, E—all isotypes of var. /ongiuscula) Stableford 124 (BM); Wilkins 
69 (BM). Nightingale: Christophersen 2090 (BM), 2090b (O), 2106 (O); Fleming 63 (E); 
Moseley 17 Oct. 1873 s.n. (K, BM -syntypes of var. /usciniae) Oct. 1873 s.n. (BM, E -syntypes of var. 
lusciniae) Stableford 3 (K), s.n. (K). Gough: Discovery ‘W.S’ Exped. s.n. (BM); Fleming 19 (E); van de 
Merwe 34 (PRE); Wace 60 (BM, PRE), 77 (BM). 

Acaena stangii Christoph. 
Endemic. First discovered by Christophersen on Tristan da Cunha in January 1938 growing 
in Empetrum heath and described by him as new (Christophersen, 1944: 7-10). He 


* Bitter designates the holotype of this variety as being in the herbarium at Stockholm but fails to cite a collection. 


360 E. W. GROVES 


considers this species to be more related to the Australian forms of Acaena than to those 
from South America. 

Tristan: Bonomi s.n. (NBG [herb. Mus. Austro-Afric. 1141]); Christophersen 46 (O, BM), 
831 (O -holotype, BM), 1271 (O, BM); Dickson 112 (BM, AAS); Mejland 632 (O), 814, (O); 
Wace 1.13 (BM), 1.69 (BM). Inaccessible: Christophersen 2360 (O), 2581 (O, BM). Gough: 
Wace 81 (BM), 107 (BM). 


Malus sylvestris subsp. mitis (Wallr.) Mansf. 

Introduced. Wace & Holdgate (1976: 45) report that on Tristan da Cunha apples have been 

widely planted around the island at low altitude but do not appear to be spreading. A few 

trees at Stony Beach (Crawford, 1941: 48) were known to have been there as least earlier than 

1937. Apples have also been reported on Inaccessible Island, where a few were found 

amongst Spartina tussock at Saltbeach and Blenden Hall (Wace & Dickson, 1963: 333). 
Tristan: Fleming 54 (E). 


Prunus persica (L.) Batsch 

Introduced. Herbarium material collected on Tristan da Cunha during the 1937-38 

Norwegian Scientific Expedition at both Sandy Point and between Hagan Point and Miller 

Point came from stunted trees, probably originating from stock planted by early settlers. 
Tristan: Christophersen 422 (O, BM); Mejland 1173 (O, BM). 


Rosa rubiginosa L. 
Probably introduced. Collected on Tristan da Cunha in 1937 in a gorge on the road to the 
Potato Patches. 

Tristan: Mejland 154 (O, BM). 


Rosa spinosissima L. 
Probably introduced. Collected on Tristan da Cunha in February 1904 but without precise 
locality. 

Tristan: Bonomi 15 (NBG [herb. Mus. Austro-Afric. 1 140]). 


Rubus saxatilis L. 
Introduced. First noted on Tristan da Cunha by Mrs Barrow, who refers to the presence of 
blackberry as being ‘scanty’ (Barrow, 1910: 177). In 1937 Christophersen collected it in a 
pasture on Hill Piece. Wace reports that blackberries now grow quite well, particularly at 
Sandy Point (Wace & Holdgate, 1976: 45). 

Tristan: Christophersen 1330 (O). 


CRASSULACEAE 


Crassula pellucida L. 
Introduced. First found on Tristan da Cunha by Stableford during the summer of 1954 by a 
stream or “watron’ near the Settlement. 

Tristan: Stableford | (BM), s.n. (K); Wace T.63 (BM). 


CALLITRICHACEAE 


Callitriche christensenii Christoph. 

Endemic. Originally described from material collected on Gough Island by Christensen in 
1933, although specimens gathered there nearly 20 years before by R. N. R. Brown during 
the visit of the Scottish National Antarctic Expedition, April 1904, have since been 
identified as belonging to this taxon. The species has also been subsequently found on 
Tristan da Cunha, Inaccessible Island and Nightingale Island. 

Tristan: Christophersen 544 (O), 1272 (O); Dickson 111 (BM, AAS); Mejland 1360 (O). 
Inaccessible: Christophersen 2388 (O), 2438 (O), 2560 (O). Nightingale: Christophersen 2093 
(O). Gough: Christensen s.n. (O -holotype); Brown s.n. (K); van der Merwe s.n. (PRE); Wace 
15 (BM), 79 (BM). 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 361 
MYRTACEAE 


Eucalyptus viminalis Labill. 
Introduced. It is not known when this species was introduced on Tristan da Cunha, but Wace 
& Holdgate (1976: 45) report that it had thrived near the pine plantation at Sandy Point prior 
to 1968. Fleming collected a specimen at this locality in December 1971. 

Tristan: Fleming 40 (E). 


ONAGRACEAE 


Oenothera indecora subsp. bonariensis Dietr. 
Introduced. Recorded on Tristan da Cunha by Stableford in 1953-54. Probably a garden 
escape. 

Tristan: Stableford s.n. (K). 


UMBELLLIFERAE 


Apium australe Thouars CELERY 

A. goughense Bak.f. & Wilk. 

Native. First recorded on Tristan da Cunha by Aubert Du Petit-Thouars in 1793 and later 
frequently collected on all islands of the group. It occurs from sea-level to about 1000 m, 
mainly on rocks and scree slopes amongst moss and dense fern. A. goughense, described by 
Baker & Wilkins (Wilkins 1925, p. 68) and based on material collected by the latter in 1922 
on Gough Island, appears to exhibit little, if any, difference in characters from some of the 
variations found in the polymorphic A. australe, and therefore I follow Christophersen 
(1968) in not regarding it as a distinct entity. 

Tristan: Bonomi 18 (K, NBG [herb. Mus. Austro-Afric. 1138]); Carmichael s.n. (BM, K); 
Christophersen 25 (BM, O) 427 (O, BM), 445 (O, BM), 635 (O, BM), 1020 (O, BM), 1276 (O), 1277 (O), 
1324 (O); Dickson 43 (BM, AAS); Dyer 3568 (PRE, NBG, K); Keyte/ 1800 (NBG, K, BM), s.n. (PRE 
[{herb. Marloth. 4723]); Mejland 1591 (O), 1636 (O, BM,) 1637 (O, BM); Moseley s.n. (K); Rogers (BM); 
Thouars, s.n. (P -holotype of A. australe); Wace T.73 (BM). Inaccessible: Christophersen 2419 (O, 
BM), 2421 (O, BM), 2422 (O), 2550 (O, BM), 2551 (O), 2587 (O, BM); Fleming s.n. (E); Moseley s.n. 
(BM. K); Stableford 126 (BM), 127 (BM). Nightingale: Christophersen 2007 (O, BM), 2045 (O), 2046 
(O, BM), 2065 (O), 2135 (O), 2182 (O, BM); Stableford 76 (BM), s.n. (K). Middle: Christophersen 2026 
(O, BM), 2027 (O), 2028 (O). Stoltenhoff: Christophersen 2149 (O). Gough: Brown s.n. (K, E); 
Christensen s.n. (O, BM); Fleming 16 (E); MacMillan s.n. (PRE); van der Merwe 45 (PRE), 58 (PRE, 
K): Wace 34 (BM) 89 (BM), 117 (BM); Wilkins 82 (BM), 83 (BM), & 84 (BM-synotype of A. 
goughense). 


Centella asiatica (L.) Urban* 
Hydrocotyle asiatica L. 
Introduced. So far only recorded from Tristan da Cunha, where the first specimen was 
collected by Bonomi in 1904. Occasionally found in short grass of gardens, pasture and 
waste ground near the Settlement, at the Potato Patches, and also near Jenny’s Watron. 
Tristan: Bonomi s.n. (NBG [herb. Austro-Afric. 1137]); Christophersen 16 (O), 269 (O); Dickson 89 
(BM, AAS); Dyer 3537 (PRE, NBG); Keytel 1799 (NBG); Mejland 105 (O) 349 (O, BM); Stableford 8 
(BM), 98 (BM), 113 (K); Wace T.54 (BM), T.90 (BM). 


Hydrocotyle capitata Thouars 

Native. First collected on Tristan da Cunha by Aubert Du Petit-Thouars in 1793. Since then 
the species has been found on the island several times, and occurs not uncommonly in 
“grassy places from the lowlands to middle altitudes’ (Christophersen, 1968: 15). It has been 
found between rocks on Stag Beach and also in a small valley bog amongst moss at Jenny’s 
Watron. It also occurs on Inaccessible and Gough Islands in similar damp situations, 
although it has not yet been noted from Nightingale Island. It has never been recorded 
growing wild outside the group. 


* A 1962 record of this species on Inaccessible Island attributed to Dickson is given in Wace & Dickson (1965, 
Appendix A), but no specimen has been traced to substantiate it. 


362 E. W. GROVES 


Tristan: Carmichael s.n. (BM, K); Christophersen 11 (O, BM), 550n (O), 1241 (O), 1560 (O); Dickson 
96 (BM, AAS); Dyer 3538 (PRE, NGB, K), 3566 (PRE, NBG, K); Keytel 1798 (NBG, K); Macgillivray 
356 (K); Mejland 107 (O), 621 (O, BM), 1152a (O, BM), 1428 (O, BM); Moseley s.n. (K); Rogers s.n. 
(K); Siggeson 45 (O); Stableford 31 (BM), 100 (BM), s.n. (K); Thouars 95 (P -holotype, BM); Wace 
T.55 (BM). Inaccessible: Christophersen 1279 (O), 2325 (O), 2579 (O); Fleming s.n. (E); Moseley s.n. 
(BM, K). Gough: Brown s.n. (K, E); Discovery ‘W.S.’ Exped. s.n. (BM); van de Merwe 3 (PRE, K); Wace 
50 (BM), 126 (BM); Wilkins 74 (BM); Christensen s.n. (O) from Gough Island probably also belongs to 
this species. 


RUBIACEAE 


Galium aparine L. 
Introduced. First found on Tristan da Cunha in December 1937 in a garden at the 
Settlement, where it occurred as a weed. 

Tristan: Christophersen 407 (O); Mejland 1294 (O, BM). 


Nertera assurgens Thouars 

Endemic. First found on Tristan da Cunha by Aubert Du Petit-Thouars in 1793, the type 
being described by him and based on his own material (Aubert Du Petit-Thouars 1808; 42). 
Since then it has been collected several times on the island, where it usually grows in moss 
in damp localities between 150 and 650 m. It has been found once on Inaccessible Island, but 
the recent record by Stableford from Nightingale Island is probably an error for Tristan da 
Cunha (see footnote). 

Tristan: Carmichael s.n. (BM, K); Christophersen 458 (O, BM, PRE), 1722 (O); Dickson 12 (BM, 
AAS); Fleming 68 (E); Keytel 1801 (NBG, K); Mejland 1634 (O, BM), 1725 (O, BM), 1731 (O); 
Stableford 17* (K), 61 (BM); Thouars s.n. (P -holotype, BM); Wace T.30 (BM). Inaccessible: Moseley 
s.n. (BM). 


Nertera depressa Banks & Sol. ex Gaertn. FOWLS’S BERRY BUSH; CHICKEN BERRY 

N. granadensis (L.f.) Druce 
Native. First found on Tristan da Cunha by Aubert Du Petit-Thouars in 1793. This small, 
creeping plant, rooting at the nodes, forms dense patches amongst grass and was in 1937 
reported by Dyer (1939: 605) as being ‘frequent on the margin of the Settlement plateau 
towards the landing beach’. It also occurs up the mountain slope, where during 
the same expedition Dyer found it ‘often with numerous little red berries’. It has also been 
found on Inaccessible, Nightingale and Gough Islands. 

Tristan: Bonomi 17 (K, NGB [herb. Mus. Austro-Afric. 1124]); Carmichael s.n. (BM, K); 
Christophersen 7 (O, BM), 62 (O), 550h (O); Dyer 3539 (PRE, NBG, K), 3581 (NBG, K); Fleming 88 
(E); Glass s.n. (BM); Mejland 136 (O), 1732 (O); Moseley s.n. (K); Rogers 16 (BM), s.n. (K); Siggeson 46 
(O)t; Stableford 62 (BM); Thouars s.n. (P); Wace T.6 (BM), T.64 (BM), T.70 (BM). Inaccessible: 
Christophersen 2549 (O, PRE). Nightingale: Christophersen 2040 (O); Fleming 65 (E), s.n. (E). Gough: 
Brown s.n. (K, E); Discovery ‘W.S.’ Exped. s.n. (BM); Fleming 32 (E); Wace 61 (BM), 111 (BM), 129 
(BM), 149 (BM). 

Christophersen (1944: 12) has described a forma fimbriata which he states as differing from the 
typical form in having the cells of the callose margins to the leaves developed into papillae and hairs. It 
is not possible to say how widespread this form may be until a critical examination can be made on all 
material collected so far from Tristan da Cunha and from the other islands. Material collected during 
the Norwegian Scientific Expedition 1937-38 which Christophersen has assigned to this form is as 
follows: 

Tristan: Mejland 344 (O), 1374 (O, BM), 1388 (O -holotype of f fimbriata BM), 1619 (O), 1632 (O, 
BM), 1633 (O), 1638 (O), 1723a (O), 1724 (O, BM), 1726 (O), 1727 (O), 1728 (O, BM), 1729 (O), 1730 
(O), 1733 (O), 1734 (O, BM), 1761 (O, BM, PRE), 1771 (O, BM), 1817 (O). 


Nertera holmboei Christoph. 
Endemic. Described as a new species by Christophersen (1944: 13-14) from a specimen 


* The label accompanying this gathering gives Nightingale Island as the locality, but most probably the specimen 


was collected on Tristan da Cunha. 
+ Not seen by me in Oslo herbarium in October 1973, but cited as being there by Christophersen (1937). 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 363 


collected by him amongst tussock in a rocky place on Middle Island, February 1938, during 
the Norwegian Scientific Expedition to Tristan da Cunha 1937-38. The earliest known 
record for the group however is a specimen collected by Moseley on Inaccessible Island on 
16 October 1873 during the HMS Challenger Expedition. He also found it the following day 
on Nightingale Island. Besides the above mentioned islands it has also been recorded on 
Stoltenhoff Island. It favours shady moist places, amongst mosses. Christophersen (loc. cit.) 
states that it is distinguished by its large, broadly ovate to reniform leaves with even margins 
and by its large fruits (up to 6 mm in diameter) with seeds up to 3.3 mm long. 

Inaccessible: Christophersen 2476 (O, BM); Dickson 132 (BM, AAS): Moseley s.n. (K). 
Nightingale: Christophersen 2103 (O, BM); Dickson 138 (BM, AAS); Moseley s.n. (K). 
Middle: Christophersen 2021 (O -holotype, BM). Stoltenhoff: Christophersen 2148 (O). 


COMPOSITAE 


Anthemis cotula L. 
Introduced. First collected on Tristan da Cunha in 1904 by Bonomi. It has been found near 
the Settlement as a weed of cultivation. 
Tristan: Bonomi 10 (NBG [herb. Mus. Austro-Aric. 1133]); Keytel 1815 (NBG, BM); Mejland 1292 
(O, BM). 
Bellis perennis L. 
Introduced. Originally collected on Tristan da Cunha by Mejland in 1937 near the Settle- 
ment. It particularly favours path-sides and trampled turf. Double forms have been noted. 
Tristan: Mejland 122 (O, BM); Siggeson 37 (O); Wace T.91 (BM). 


Chevreulia sarmentosa (Pers.) Blake 

C. stolonifera Cass.; Xeranthemum caespitosum Thouars 
Native. First collected on Tristan da Cunha in January 1793 by Aubert Du Petit-Thouars 
whose material became the type of his Yeranthemum caespitosum. It has been found on the 
island on several occasions, growing in both pasture and meadow and also in the short turf 
near the cliff edge west of the Settlement. The species has also been recorded for Inaccessible 
Island during the Norwegian Expedition of 1937-38. 

Tristan: Bonomi 8 (NGB [herb. Mus. Austro-Afric. 1134], K); Carmichael s.n. (BM, K); 
Christophersen 1129 (O): Dickson 39 (BM, AAS); Mejland 1104 (O, BM), 1165 (O), 1664 
(O); Siggeson 52 (O); Stableford 94 (BM); Thouars s.n. (P* -holotype of X. caespitosum, 
BM); Wace T.84 (BM). Inaccessible: Christophersen 2575 (O, BM). 


Conyza sumatrensis (Retz.) E. H. Walker 

C. floribunda Kunth 
Introduced. First collected on Tristan da Cunha by Mejland near the Settlement in 1938, and 
since found several times on the island in a few other localities. 

Tristan:t Christophersen 1787 (O); Dickson 14 (BM), 178 (BM); Mejland 1296 (O), 1476 
(O, BM), 1645 (O, BM); Stableford 2 (BM), s.n. (K). 

Cotula australis (Sieber ex Sprengel) J. D. Hook. 

Introduced. First found on Tristan da Cunha by both Macgillivray and Milne during the visit 
of HMS Herald in November 1852. It has been collected on the island on several occasions 
since, as well as on Nightingale Island by Stableford in 1954. It is a plant of open grassy 
places. 

Tristan: Christophersen 1130 (O, BM); Dickson 34 (AAS), 86 (BM); Dyer 3528 (PRE, BM); 
Macgillivray s.n. (K); Mejland 108 (O), 126a (O), 1308 (O); Milne s.n. (K); Stableford 6 
(BM), 101 (BM), s.n. (K); Wace T.60 (BM). Nightingale: Stableford 74 (BM), s.n. (K). 

Cotula goughensis R. N. R. Brown 
Endemic. Recognised as a new species by R. N. R. Brown (1905: 242, plate 9) from material 
* Not located by me at P in September 1975. 


t+ Some of the Tristan gatherings have been identified as Conyza bonariensis (L.) Cronq., but all material I have 
seen has proved to be C. sumatrensis. 


364 E. W. GROVES 


he collected on Gough Island during the visit of the Scottish National Antarctic Expedition 
in April 1904. In later visits by Wilkins in 1922 and by Wace in 1955 the species was found 
to be common and widely dominant at lower levels, decreasing with altitude. 

Gough: Brown s.n. (K, E—both syntypes); Christensen s.n. (O); Fleming 37 (E); van der 
Merwe 31 (PRE); Wace 3 (BM); Wilkins 78 (BM), 79 (BM). 


Cotula moseleyi Hemsley 
Endemic. Reported only from Nightingale Island (Fig. 20), where it was first found by 
Moseley during the visit of HMS Challenger in October 1873. It grows abundantly in damp 
meadows and tussock grassland. 

Nightingale: Christophersen 2104 (O, BM), 2168 (O, BM), 2236 (O); Dickson 140 (BM); Fleming 59 
(E); Moseley s.n. (K -holotype, BM); Stableford 75 (BM), s.n. (K). 


Crepis capillaris (L.) Wallr. 
Introduced. First collected on Tristan da Cunha near the Settlement by Dickson in February 
1962. 

Tristan: Dickson 106 (BM). 


Gnaphalium candidissimum Lam. 

Introduced. This species, amongst others, is stated in a recent paper by Wace & Holdgate 
(1976: 45) as being ‘characteristic of waste ground in and around the Settlement’ on Tristan 
da Cunha. No herbarium material for G. candidissimum has been seen from the island, 
although possibly it was collected there by Wace during his 1968 visit. 


Gnaphalium luteo-album L. MUCK WEED 
Introduced. First collected on Tristan da Cunha by Bonomi 1904. Common on lowland 
pastures and on the mountain slopes up to 1220 m. It has also been found on Inaccessible 
Island. 

Tristan: Bonomi 22 (NBG [herb. Mus. Austro-Afric. 1136]); Christophersen 8 (O, BM); 
Dyer 3545 (PRE, NBG); Keytel 1803 (NBG, BM): Mejland 152 (O, BM), 1662 (O, BM), 1663 (O, BM); 
Siggeson 38 (O); Stableford 18 (BM), 95 (BM), 96 (BM), s.n. (K); Wace T.53 (BM). Inaccessible: 
Christophersen 2420 (O). 


Gnaphalium purpureum L. 
Introduced. First collected on Tristan da Cunha by Wace in 1955, occurring in cultivated 
ground around the Settlement and occasionally on lower slopes. 

Tristan: Dickson 107 (BM): Wace T.61 (BM). 


Gnaphalium thouarsii Sprengel Cow PUDDING GRASS 

G. pyramidale Thouars, non Berg. 

Endemic. First collected on Tristan da Cunha by Aubert Du Petit-Thouars in January 1793 
and subsequently also on Inaccessible, Nightingale and Gough Islands. It commonly occurs 
in wet stream gullies and on open grassy areas. 

Tristan: Carmichael s.n. (BM); Christophersen 420 (O), 513 (O), 547 (O, BM), 1329 (O):; Fleming 70 
(E); Mejland 612 (O), 643 (O, BM), 1161 (O), 1669 (O, BM), 1670 (O, BM); Milne s.n. (K); Moseley s.n. 
(K); Rogers 18 (BM); Stableford s.n. (K); Thouars s.n. (P -holotype, BM). Inaccessible: Christophersen 
2358 (O, BM), 2400b (O), 2616 (O); Moseley s.n. (K, BM); Stableford 128 (BM); Wilkins 67 
(BM). Nightingale: Christophersen 2041 (O, BM), 2212 (O); Stableford s.n. (K). Gough: 
Brown s.n. (K, E); Discovery ‘W.S.’ Exped. s.n. (BM); Fleming 3 (E); van der Merwe 39 (K); 
Wace 14(BM), 69 (BM), 86 (BM); Wilkins 85 (BM). 


Hypochoeris glabra L. 
Introduced. Reported only from Tristan da Cunha, where the earliest collection was made by 
Moseley in October 1873. It has been found as a ruderal of cultivated ground and in short 
turf near the Settlement, and also at the Potato Patches. 

Tristan: Dickson 40 (BM); Keytel 1818 (NBG); Mejland 308 (O, BM), 348 (O), 1010 (O, 
BM); Moseley s.n. (BM, K); Stableford 27 (BM), s.n. (K); Wace T.57 (BM). 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 365 


AY ' mp 
~ *« } ¢ 7 = ce | 
ee , 4 
1, ss x “ij 
* A. 


ne ba r 
. "4 : i 
je ‘ 
i oe 
~* ¢ » ’ 


4 
i 2 
‘ ‘4 y >, 


=” Naat! ee 
Fig. 20 N. M. Wace collecting Cotula moseleyi and Apium australe amid shoulder-high tussock 


grassland above Petrel Bay, north side of Nightingale Island. Middle Island, with Stoltenhoff 
Island beyond, form the background. November 1976. Photograph: C. D. Ollier. 


366 E. W. GROVES 


Lactuca serriola L. 

L. scariola L. 
Introduced. This weed of cultivation was noted by Aubert Du Petit-Thouars on Tristan da 
Cunha in 1793 and reported in his list (Aubert Du Petit-Thouars, 1808: 46) under the name 
L. sativa L. It has not, as far is known, been subsequently recorded from the island. 

Tristan: Thouars s.n.* 


Lagenophora nudicaulis (Commerson) Dusén 

L. commersonii Cass. 
Native. First found on Tristan da Cunha in 1793 by Aubert Du Petit-Thouars who recorded 
it in his list (Aubert Du Petit-Thouars, 1808: 40) under the name Calendula pusilla. It grows 
in damp boggy places particularly amongst Sphagnum. 

Tristan: Christophersen 24 (O), 49 (O), 76 (O, PRE) 550m (O); Dickson 31 (BM, AAS); 
Dyer 3569 (PRE); Mejland 623 (O), 1175 (O); Stableford 2+ (K); Thouars s.n. (P). Inacces- 
sible: Christophersen 2491 (O). Gough: Wace 73 (BM), 119 (BM). 


Leontodon taraxacoides (Vill.) Mérat 
Introduced. This species was collected on Tristan da Cunha during 1955-56. 
Tristan: Stableford 97 (BM). 


Leucanthemum vulgare Lamb. 
Chrysanthemum leucanthemum L. 
Introduced. First collected on Tristan da Cunha by Bonomi in February 1904. The species 
frequently occurs not only in the lowland pasture, but also on semi-stabilised ground, 
boulder scree, and in rock crevices. It has also been found as a weed at the Potato Patches. 
Tristan: Bonomi 9 (NBG [herb. Mus. Austro-Afric. 1135]); Dyer 3544 (PRE, NBG); Keytel 
1816 (NBG, BM); Mejland 350 (O, BM); 1436 (O, BM), 1452 (O, BM), 1646 (O, BM), Rogers 9 (BM); 
Stableford 30 (BM), s.n. (K), Wace T.23 (BM). 


Senecio vulgaris L. 
Introduced. This species has been found as a weed on Tristan da Cunha both at the Settle- 
ment and at the Potato Patches. It was first collected on the island by Moseley in October 
1873. 

Tristan: Christophersen 404 (O, BM); Dyer 3527 (PRE, NBG); Keytel 1814 (NBG, BM); Mejland 128 
(O), 170 (O); Moseley s.n. (BM); Stableford 28 (BM), s.n. (K). 


Sonchus asper (L.) Hill 
Introduced. First found on Gough Island by van der Merwe in January 1957, when he stated 
on his label to the herbarium specimen that it was ‘common inland’. It has been subse- 
quently found on Tristan by Dickson near the Settlement in February 1962. 

Tristan: Dickson 104 (BM). Gough: van der Merwe 49 (PRE). 


Sonchus oleraceus L.+ 

Introduced. The earliest collection was made on Tristan da Cunha by Carmichael in 1817; in 
1962 Dickson noted that it was common throughout the island. It has been found on all six 
islands of the group. 

Tristan: Bonomi s.n. (NBG [herb. Mus. Austro-Afric. 1132]); Carmichael s.n. (BM); 
Christophersen 3 (O); Dickson 105 (BM); Keytel 1802 (NBG, K); Mejland 114 (O, BM), 132 
(O, BM), 172 (O), 1150 (O, BM), 1554 (O, BM); Moseley s.n. (K); Stableford s.n. (K); Wace 
T.89 (BM). Inaccessible: Christophersen 2617 (O); Fleming s.n. (E); Moseley s.n. (BM, K). 
Nightingale: Christophersen 2107 (O), 2109 (O). Middle: Christophersen 2023 (O). Stoltenhoff: 
Christophersen 2150 (O). Gough: Brown s.n. (K, E—wrongly labelled as Hypochaeris glabra); 
Christensen s.n. (O); Wace 10 (BM). 

* No specimen could be found when I searched for it in the Paris Herbarium in September 1975. 

{ The label on the Kew sheet gives the locality as Nightingale, but this is probably in error for Tristan (see remarks 
under Stableford, p. 352). 

7 There exists in K a gathering of another Sonchus collected by R. N. R. Brown during the visit of the Scottish 
National Antarctic Expedition to Gough in 1912 which may be S. asper or even S. arvensis, but as the single 
specimen is juvenile and sterile its identity cannot be determined. 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 367 


CAMPANULACEAE 


Lobelia erinus L. 
Introduced. The earliest collection is that of Stableford in 1954 on Tristan da Cunha where it 
is now a common weed of open cultivated ground around the Settlement and Mission 
Garden. 

Tristan: Stableford 13 (BM), s.n. (K); Wacé T.56 (BM). 


EMPETRACEAE 


Empetrum rubrum Vahl TRISTAN CRANBERRY 

E. nigrum var. rubrum (Vahl) Hemsley; E. medium Carmich. 

Native. First collected on Tristan da Cunha by Aubert Du Petit-Thouars in 1793 and since 
then found on all the islands of the group except Middle Island, although in all probability it 
also occurs there. It is commonly found associated with in the Blechnum/Phylica heath on 
the lower slopes and as a creeping ecotype on upper scree slopes. Fascinated specimens are 
not uncommon. In addition to the type, two other forms and one variety have 
been recognized from the Tristan-Gough group by some workers: the forma medium 
(Carmich.) Good (Good, 1927: 517) has narrower leaves with distinctly apiculate tips; the 
forma pulvinatum Christoph. (Christophersen, 1944: 11-12, and Christophersen, 1968: 19) 
being of a cushion-like habit and has glabrous leaves, only a few of which can be 
characterized as subspathulate; and the var. tristanitorum Christoph. (Christophersen, 1944: 
10-11, and Christophersen, 1968: 18) distinguished by ‘its prostrate matted habit, the shape 
of its long, narrowly elliptic leaves, lanceolate or oblanceolate rarely truly elongate, and 
large, light-red, shining translucent berries, sweet to the taste’ (cf. the smaller sour-tasting 
fruit of the type). This variety tristanitorum is said to grow at high altitudes (from about 
1000 m. up to the Peak) and is relished by the Tristanites who call the fruits ‘Peak Berries’ 
(Christophersen, 1968: 18). However, it seems that these forms and the variety are of 
doubtful validity, as E. rubrum from the islands, as elsewhere, exhibits a wide degree of 
polymorphism. 

Tristan: Bonomi 19 (PRE), s.n. (NBG [herb. Mus. Austro-Afric. 1125]); Christophersen | 
(O, BM), 414 (O, BM), 530 (O, BM), 1031 (O); Dickson 121 (BM, AAS); Dyer 3520 (PRE), 
3548 (PRE, NBG), 3567 (PRE, NBG); Fleming 79 (E), 84 (E), 92; (E) Keytel s.n. (PRE [herb. 
Marloth 4722]); Mejland 1011 (O, BM), 1012 (O, BM), 1138 (O); Rogers 8 (BM); Siggeson 
39 (O); Stableford 56 (BM), 99, (BM); Thouars s.n. (P); January 1877 [no collector] 15 (BM). 
Inaccessible: Christophersen 2497 (O, BM), 2546 (O, BM), 2597 (O); Wilkins 42b (BM). 
Nightingale: Christophersen 2033 (O, BM), 2095 (O, BM); Moseley s.n. (BM). Stoltenhoff: 
Christophersen 2153 (O). Gough: Brown s.n. (E); Fleming 5 (E), 34 (E); van der Merwe 35 
(PRE); Wace 35 (BM). 

Herbarium material recognised by certain workers as belonging to the forms and variety 
mentioned above is as follows: ‘ 


forma medium (Carmich.) Good 
Tristan: Carmichael s.n. (BM -holotype); Macgillivray s.n. (BM); Moseley s.n. (BM); Wace 
T.43 (BM). Gough: Wace 71 (BM). 


forma pulvinatum Christoph. 

Tristan: Christophersen 31 (O -holotype, BM), 630 (O), 1128 (O), 1380 (O); Mejland 1715 
(O). 
var. tristanitorum Christoph. 


Tristan: Christophersen 628 (O, BM), 629 (O), 1745 (O -holotype, BM), 1779 (O); Mejland 
1359 (O); Wace T.47 (BM). Gough: Wace 109 (BM), 125 (BM). 


PRIMULACEAE 


Anagallis arvenis L. 
Introduced. First cited (but without reference to a collector) for Tristan da Cunha by 
Hemsley (1885, Pt.II: 145) as being ‘among introduced plants more or less established’. The 


368 E. W. GROVES 


part of his account relating to this island (loc. cit.) was based on the collection of Moseley, 
together with the earlier material of Carmichael and Macgillivray. No specimen has been 
traced in any of the extant specimens of these collectors in the herbaria at BM, K or E. The 
earliest gathering seen, made by Keytel in 1908-09, exists in K. Anagallis arvensis has now 
become frequent as a weed of cultivation and occurs in waste places, both around the 
Settlement and at the Potato Patches. 

Tristan: Christophersen 287 (BM, O); Keytel 1811 (NBG, K); Dickson 108 (BM); Mejland 304 (O); 
Stableford 16 (BM), s.n. (K). 


BORAGINACEAE 


Lithospermum sp. 
Introduced. Reported for Tristan da Cunha by Christophersen (1937: 10), based on a single 
gathering made by Siggeson in November 1934. Christophersen (loc. cit.) states that he was 
‘unable to match it with any known species of this genus’. No further material has since been 
discovered on the island. 

Tristan: Siggeson 48 (O)*. 


Myosotis discolor Pers. 

Introduced. First collected on Tristan da Cunha in 1937. It has been found near the 

Settlement, where in 1955 it was noted to be common in pastures, and at the Potato Patches. 
Tristan: Christophersen 403 (O, BM); Dickson 84 (BM); Mejland 125 (BM), 128b (O), 176 (O, BM); 

Stableford 12 (BM), 107 (BM), 118 (BM), s.n. (K); Wace T.58 (BM). 


CONVOLVULACEAE 


Calystegia sepium subsp. americana (Sims) Brummittt 
Native. First collected on Tristan da Cunha in 1908-09 by Keytel who remarked that it was 
‘not uncommon’. Since that time it has been found at several localities on that island, e.g. 
around the Settlement, and at Stony Beach, Sandy Point, Anchorstock Bay and at the Potato 
Patches. It has also been collected on Inaccessible Island. The comparatively large corollas 
(50-60 mm long) are pale pink or violet-pink in colour. 

Tristan: Christophersen 299 (O, BM), 447 (O, BM); Mejland 1135 (O, BM), 1407 (O), 
1413 (O, BM); Keytel 1809 (NBG, K, BM); Stableford 26 (BM), s.n. (K). Inaccessible: Skjelten in 
Christophersen 2346 (O, BM); Christophersen 2576 (O, BM), 2599 (O, BM); Dickson 131 (BM). 


Calystegia tuguriorum (G. Forster) R.Br. ex J. D. Hook. 

Probably native. A specimen collected on Inaccessible Island by Christophersen in February 
1938, during the Norwegian Scientific Expedition of 1937-38, has recently been recognized 
as this well known amphi-Pacific species (Brummitt & Groves, 1981). Unfortunately the 
specimen collected was sterile, so material with flowers and fruit is desired. The corollas of 
C. tuguriorum, which may be white or pink, are usually smaller than those of C. sepium 
subsp. americana, being 25-35 mm long. C. tuguriorum has a distribution from New 
Zealand, through Chatham Island and Masafuera (Juan Fernandez islands) to the Valdivia 
and Chiloe provinces of Chile. 

Tristan: Christophersen 2147 (BM). 


Calystegia soldanella R.Br. 

Native. First found on Tristan da Cunha in 1816-17 by Carmichael, who remarks that it 
occurs on the south-east side of the island, growing in the sand close to the shore, restricted to 
a single area. It appeared to be of recent introduction (Carmichael, 1819: 505). Unfortu- 
nately no material can be traced of any specimen gathered by him at this locality. A sheet 


* The herbarium material of this gathering was not traced in O during my visit in October 1973. 


t+ Moseley (1874: 380) mentions that on Inaccessible Island he observed a convolvulus that was ‘very abundant on 
the cultivated ground near their hut’ [of two Germans] (the brothers Stoltenhoff who had been there for two years 
fur-sealing—see p. 341 under Stoltenhoff Island). There is no herbarium material of this plant in Moseley’s 
collection at K, but the plant is more likely to have been a Calystegia than a Convolvulus sens. str. 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 369 


collected nearly a hundred years later by Bonomi in February 1904 exists in NBG, and since 
then the species has been found on several occasions, all on fixed sand and mostly on the 
island’s beaches. 

Tristan: Bonomi 24 (K), s.n. (NBG [herb. Austro-Afric. 18626]); Dickson 5 (BM); Keytel 
1808 (NBG, BM); Mejland 164 (O, BM), 1408 (O, BM); Stableford 120 (BM). 


SOLANACEAE 


Physalis peruviana L. 

Introduced. The inclusion of this taxon in the alien flora of the Tristan-Gough group is based 

on a single record of a specimen found by Christophersen on Inaccessible Island in 1938. 
Inaccessible: Christophersen 2598 (O, BM, K). 


Solanum nigrum L. TRISTAN BLACKBERRY 
Introduced. First collected on Tristan da Cunha by Keytel 1909-09, and since then it has 
become a common weed around the Settlement. It has also been found on Inaccessible Island 
where it was gathered during the visit of the Norwegian Scientific Expedition in Febrauary 
1938. Christophersen (1968: 21) remarks that ‘some of the plants are distinctly woody at the 
base, as is frequently the case in warmer regions’. 

Tristan: Dickson 103 (BM, AAS); Keytel 1807 (NBG, BM); Mejland 1291 (O); Rogers 14 (BM); 
Stableford s.n. (K). Inaccessible: Christophersen 2302 (O, K), 2324 (O, BM), 2418 (O). 


SCROPHULARIACEAE 


Verbascum virgatum Stokes 
Introduced. First collected on Tristan da Cunha by Keytel in 1908-09. Since that date it has 
been found several times, mainly around the Settlement near gardens. 

Tristan: Christophersen 1741 (O), 1783 (O); Dickson 83 (BM); Mejland 1300 (O), 1405 (O), 1451 (O, 
BM); Keytel 1820 (NBG, K). 
Veronica agrestis L. 
Introduced. Originally collected by Mejland on Tristan da Cunha in March 1938. It occurs 
at the Settlement by paths and as a weed of gardens. 

Tristan: Dickson 73 (BM), 145 (BM); Mejland 1435 (O, BM). 


Veronica serpyllifolia L. 
Introduced. First collected on Tristan da Cunha by Keytel in 1908-09. It occurs commonly 
both in damp places all over the island and in pastures near the Settlement. Dyer (1939: 605) 
notes that the species was ‘occasionally a pioneer colonizer along with Plantago lanceolata 
and Cynodon dactylon’. 

Tristan: Dickson 38 (BM, AAS); Dyer 3536 (PRE, NBG, K); Keytel 1822 (NBG, K); Mejland 135 (O, 
BM), 1631 (O, BM); Stableford 18 (K*), 29 (BM), 60 (BM), 106 (BM), s.n. (K); Wace T.52 (BM). 


VERBENACEAE 


Verbena officinalis L. 
Introduced. First collected on Tristan da Cunha by Bonomi in 1904. It occurs at the 
Settlement near gardens and in the pastures. 

Tristan: Bonomi 21 (NBG [herb. Mus. Austro-Afric. 1130]); Dickson 47 (BM); Keytel 1821 (NBG, 
BM); Mejland 341 (O, BM); Stableford 3 (BM), 24 (BM), s.n. (K). 


LABIATAE 


Leonotis leonurus (L.) R.Br. 
Introduced. This record is based on a single specimen collected on Tristan da Cunha near the 
Settlement in February 1938. It was probably an escape from a garden, perhaps originally as 
an introduction from South Africa. 

Tristan: Mejland 1644 (O, BM). 


* Localised as from Nightingale Island but probably an error for Tristan da Cunha. 


370 E. W. GROVES 


Prunella vulgaris L. 
Introduced. First found on Tristan da Cunha by Mejland in January 1938. It occurs as a 
weed of gardens at the Settlement, and also in damp situations at the Potato Patches. 

Tristan: Mejland 1287 (O, BM), 1745 (O); Stableford 19 (BM), s.n. (K). 


PLANTAGINACEAE 


Plantago lanceolata L. 

Introduced. First collected on Tristan da Cunha by Bonomi in February 1904. It is found in 
pastures around the Settlement and at the Potato Patches. Christophersen (1968: 21) remarks 
that during the Norwegian Scientific Expedition of 1937-38 the species was exceedingly 
common in the pastures of the Settlement and, in certain places, as on Hill Piece, it was the 
dominant plant in extensive meadows. More recently, during 1956-57, it was collected on 
Gough Island. 

Tristan: Bonomi s.n. (NBG [herb. Mus. Austro-Afric. 1131]); Christophersen 133la (QO); 
Dyer 3546 (BM); Keytel 1810 (NBG, BM); Mejland 159 (O), 1303 (O, BM), 1655 (O, BM), 1656 
(O, BM); Rogers 17 (BM); Siggeson 43 (O); Stableford 15 (BM), 103 (BM), 113 (BM), 193 
(K); Wace T.83 (BM). Gough: van der Merwe s.n. (PRE [Nat. Herb. Pretoria 36478)). 

Plantago major L. : 

Introduced. First collected on Gough Island by R. N. R. Brown in 1904, and later (between 
1922-25) on Tristan da Cunha near the Settlement by the Rev. & Mrs H. M. Rogers. It has 
since been found on Inaccessible Island. 

Tristan: Dyer 3531 (PRE); Mejland 1298 (O, BM); Rogers 15 (BM). Inaccessible: 


Christophersen 2341 (O, BM); Stableford 125 (BM). Gough: Brown s.n. (E); van der Merwe 
70 (PRE); Wace 66 (BM). 


CHENOPODIACEAE 


Atriplex plebeja Carmich. 
Endemic. First collected on Tristan da Cunha in 1816-17, by Carmichael who described it as 
a new species (Carmichael, 1819: 508). It was not collected again from the group until 1938, 
when during the Norwegian Scientific Expedition it was found on Nightingale, Middle and 
Stoltenhoff Islands. A specimen collected on Tristan da Cunha during the Royal Society 
Expedition of 1962 (Dickson 66) may possibly be this species, but the material is immature. 
Tristan: Carmichael s.n. (K -holotype); Dickson 66 (BM possibly this species—see above). 
Nightingale: Christophersen 2108 (OQ); Stableford 77 (BM), s.n. (K). Middle: Christophersen 2019 (O). 
Stoltenhoff: Christophersen 2151 (O). 
Chenopodium album L. 
Introduced. First recorded on Tristan da Cunha in January 1793 by Aubert Du 
Petit-Thouars. Although, in the list of plants found during the course of his visit to the island 
he remarks ‘J’ai trouvé quelques pieds de cette plante européene’ (Aubert Du Petit-Thouars, 
1808: 87), I failed to trace any specimen in P during my visit in October 1975. It does not 
appear to have been found again on the island by any subsequent visitor. 


Chenopodium ambrosioides ISLAND TEA 
subsp. ambrosioides var. tomentosum (Thouars) Aellen 
C. tomentosum Thouars 

Endemic variety. First found on Trista da Cunha by Aubert Du Petit-Thouars in January 
1793, who described it as a new species (Aubert Du Petit-Thouars, 1808: 38); Aellen (in 
Christophersen, 1968: 6) considered that it was only a variety of C. ambrosioides subsp. 
ambrosioides. The plant has since been found on Inaccessible, Nightingale and Gough 
Islands. Aubert Du Petit-Thouars (loc. cit.) remarked that it was ‘common everywhere’ on 
Tristan da Cunha. and it must have still been present in quantity 90 years later for Hemsley 
(1885: 153) to note it as being abundant. However, continual use of this plant 
over the next 70 years as a substitute tea seems to have greatly diminished its occurrence. 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 371 


The only locality on Tristan where Wace in 1955 noted it in any quantity was on the beach 
at Sandy Point. On Gough Island it grows particularly vigorously on nitrogen-rich soil 
around penguin rookeries. 

Tristan: Bonomi 11 (NBG [herb. Mus. Austro-Afric. 1126]); Carmichael s.n. (BM, P); 
Christophersen 443 (O, BM), 1746 (O, BM); Dickson 67 (BM); Glass s.n. (BM); Keytel 1804 (NBG); 
McCraken s.n. (BM); Macgillivray 354 (K), s.n. (BM); MacMillan s.n. (PRE); Milne s.n. (K); Moseley 
s.n. (BM); Rogers 13 (BM), s.n. (K); Stableford 25 (BM), s.n. (K); Thouars s.n. (P -holotype, BM); Wace 
T.82 (BM). Inaccessible: Christophersen 2385 (O, BM), 2593 (O), 2620 (O, BM); Moseley s.n. (BM). 
Nightingale: Christophersen 2063* (O). Gough: van der Merwe 25 (PRE—2 sheets collected at different 
dates); Wace 92 (BM), 151 (BM), Wilkins 76 (BM), 77 (BM). 


Chenopodium murale L. 
Introduced. Originally found on Tristan da Cunha by Keytel in 1908-09 around the 
Settlement, but not collected since. The Kew sheet of Keytel’s gathering was examined in 
1929 by Paul Aellen, who thought it represented an untypical example of his var. 
acutidentatum, described originally from South Africa (Aellen 1928: 343). 

Tristan: Keyte/ 1817 (NBG, K, BM). 


POLYGONACEAE 


Polygonum aviculare L. 
Introduced. First found on Tristan da Cunha in 1937 by Dyer, who, on his label to the 
herbarium specimen, comments that the species was ‘probably only recently introduced’. It 
is now acommon weed around the Settlement and at the Potato Patches. 

Tristan: Christophersen 290 (O); Dickson 37 (BM); Dyer 3529 (PRE, K); Mejland 1289 (O, 
BM); Stableford 9 (BM), s.n. (K). 


Rumex angiocarpus Murb. SORREL DOCK; SOUR GRASS 

R. acetosella subsp. angiocarpus (Murb.) Murb. 

Introduced, now naturalised. First collected on Tristan by Macgillivray and Milne in 1852. 
From a study of pollen in bog peat Hafsten suggests it may have invaded Tristan da Cunha 
about the same time as the discovery of the island in the early 16th century, some three 
centuries before settlement in 1810 (Hafsten, 1951: 9 and 1960: 524). This diploid form of R. 
acetosella covers extensive meadows, often in pure stands both around the Base and on the 
slopes of the Peak up to 1100 m. Wace (1967: 53) considers its spread coincided with the 
increase in numbers of sheep on the Peak, and possibly also with that of goats and rabbits. It 
has also been found on Inaccessible Island. 

Tristan: Bonomi 23 (NBG [herb. Mus. Austro-Afric. 1127]); Christophersen 298 (O, BM); 
Dyer 3519 (PRE, K), 3580 (PRE, NBG, K); Fleming 93 (E); Keytel 1806 (NBG, K, BM); 
Macgillivray 353a (K), 353b (K), s.n. (BM, K); Mejland 1545 (O, PRE), 1771la (O); Milne s.n. 
(K); Moseley s.n. (BM, E); Saunders s.n. (K); Siggeson 42 (O); Stableford 102 (BM), 112 
(BM), s.n. (K); Wace T.45 (BM), T.59 (BM), T.67 (BM). Imaccessible: Moseley s.n. (K). 
Gough: Wace (1967: 53) reports a single plant of R. angiocarpus on Gough Island which 
inadvertently ‘had been imported with some vegetables’. The plant was destroyed to prevent 
the possibility of it eventually invading upland areas, in the same manner as it has spread on 
Tristan da Cunha over the years. 


Rumex crispus L. 
Introduced. First found on Tristan da Cunha by Carmichael in 1817 but apparently not 
collected again since that date. 

Tristan: Carmichael s.n. (BM). 


Rumex frutescens Thouars PIG DOCK 


R. cuneifolius Campd. 
Native. First collected on Tristan da Cunha in 1793 by Aubert Du Petit-Thouars who 


described his material as a new species Rumex frutescens. Rechinger (1954: 62) stated that he 
could find no characters which separate R. frutescens (hitherto regarded as endemic to 
Tristan) from the polymorphic R. cuneifolius of South America, and in his treatment of the 


B92 E. W. GROVES 


South African species gave the former name in synonymy of the latter (loc. cit.: 61). 
However as R. frutescens was published eight years earlier it is therefore the valid name for 
the taxon, and was correctly given as such by the same author in Flora Europaea 1: 85 § 22 
(1964). On Tristan this species is a common lowland plant of beaches, waste ground and 
damp places; it also occurs on cinder scree slopes. R. frutescens has also been found on 
Inaccessible and Gough Islands, on the latter frequently on disturbed soil around penguin 
rookeries. As the fruits of R. frutescens are very buoyant, Rechinger (1954: 63) considers that 
its arrival in the group may well have been due to ocean currents drifting the seed from 
beaches along the South American coastline, where the species commonly occurs. 

Tristan: Bonomi 12 (NBG [herb. Mus. Austro-Afric. 1129], BM); Carmichael s.n. (BM); 
Christophersen 6(O), 14 (O, BM, K, PRE), 829 (O, BM); Dickson 2 (BM); Dyer 3518 (PRE, NBG); 
Moseley s.n. (BM, E); Thouars s.n. (P -holotype); Wace T.93 (BM). Inaccessible: Christophersen 2573 
(O, BM, K); Fleming s.n. (E). Gough: Brown s.n. (E); Wace 93 (BM). 

Rumex obtusifolius L. subsp. obtusifolius 

R. obtusifolius var. agrestis Fries; R. obtusifolius subsp. agrestis (Fries) Celak 
Introduced. First found on Tristan da Cunha by Bonomi in 1904. Several subspecies of R. 
obtusifolius have been described by various authors but, according to J. E. Lousley (pers. 
comm., 1975), only subsp. obtusifolius has been recorded from the Tristan-Gough islands. 
On Tristan the species occurs as a ruderal around the Settlement and at the Potato Patches, 
whilst on Inaccessible and Gough Islands it has been found on talus slopes, beaches and 
around elephant seal wallows and penguin rookeries. 

Tristan: Bonomi s.n. (NBG [herb. Mus. Austro-Afric. 1128]: Christophersen 297 (O, BM), 902 (O, 
BM); Keytel 1813 (NBG, K); Mejland 115 (O, BM), 169 (O, BM): Stableford s.n. (K). Inaccessible: 
Christophersen 2596 (O), 2618 (O). Gough: Brown s.n. (K); Christensen s.n. (O—probably this species); 
MacMillan s.n. (PRE); van der Merwe 63 (PRE); Wace 5 (BM). 


Rumex steudelii Hochst. 

Introduced. Represented on Tristan by material gathered on a single occasion in a grass field 
at the Settlement by Mejland in January 1938. Regarding this collection, J. E. Lousley 
(pers.comm., 1975) provided the following remarks: ‘The material collected by Mejland 
under No. 610 certainly belongs to the subsection hamati, with sharply recurved teeth on the 
tepals, and appears to be R. steudelii Hochst., which is a well known Abyssinian species. 
Rechinger in his monograph on the Rumex of Africa lists specimens from South Africa 
which he says are allied to R. steudelii, but unfortunately does not state clearly how they 
differ from the typical form’. Lousley further commented that the spread of this species in 
South Africa and also on Tristan, away from its main centre of distribution in the highlands 
of Ethiopia, Somalia and South-West Arabia, was probably due to the fact that seeds of 
members of the subsection hamati are easily distributed by animals and humans on account 
of their hooked teeth, which readily adhere to both fur and clothing. 

Tristan: Mejland 610 (O, BM). 


PIPERACEAE 


Peperomia berteroana Miq. PEPPER TREE 

P. tristanensis Christoph. 
Possibly native. The presence of this species had been known for some time before the exact 
locality on Inaccessible Island was shown to E. Christophersen during the Norwegian 
Scientific Expedition, 1937-38. After examination of the material he collected there on 2 
March 1938, Christophersen (1944: 5) considered it a new species, P. tristanensis. Later, 
however, on the basis of evidence put forward by Skottsberg (1946: 251) who compared it 
with the plant from the Juan Fernandez Islands, Christophersen (1968: 4) concluded that 
Skottsberg had established the fact that P. tristanensis was identical with P. berteroana. Itisa 
large plant, easily recognised by its fleshy stems and dark green, glossy leaves, and was found 
at an altitude of 150 m above the waterfall at Salt Beach. 

Inaccessible: Christophersen 426 (O), 502 (O), 2592 (O -holotype of P. tristanensis): Glass in 
Christophersen 1278 (O). 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 373 
EUPHORBIACEAE 


Euphorbia peplus L. 
Introduced. A few plants were found by J. H. Dickson during the Royal Society Expedition 
to Tristan da Cunha, February to March 1962, in a garden at the west end of the Settlement. 
Obviously a weed of cultivation. 

Tristan: Dickson 6 (BM). 


MORACEAE 


Ficus carica L. FIG 
Introduced. A specimen was collected in a pasture at Little Sandy Gulch on Tristan da 
Cunha by Mejland in February 1938, no doubt representing a relict of past cultivation. 
Christophersen (1968: 5) states that the shoots were sterile, arising from roots that had been 
constantly cropped close by grazing animals. 

Tristan: Mejland 1434 (O). 


SALICACEAE 


Salix babylonica L. 
Introduced. Willows, probably of this species, were first mentioned in 1910 by Barrow (1910, 
App.: 275-277), who recorded a ‘few trees on the Settlement only’ on Tristan da Cunha. It 
has since spread to a few other localities and has also been found on Inaccessible Island. The 
wood of willows is used by the Tristanites for making boat ribs. 

Tristan: Christophersen 450 (O); Dickson 48 (BM), 181 (BM). Inaccessible: Christophersen 2416 (O). 


IRIDACEAE 


Romulea rosea var. australis (Ewart) De Vos 
Introduced. First found on Tristan da Cunha by Keytel, 1908-09, an introduction probably 
from South Africa. Since then it has become naturalised along roadsides and in pasture 
around the Settlement. 

Tristan: Keyte/ 1819 (NBG, K): Stableford 10 (BM), s.n. (K). 


LILIACEAE 


Phormium tenax J. R. & G. Forster New ZEALAND FLAX 
Introduced. First mentioned on Tristan da Cunha by Barrow (1910, Appendix: 275-277) 
and now locally (although extensively) cultivated in gardens near the Settlement and at the 
Potato Patches, being used as a wind-break (Fig. 5) and for the production of rope and string. 
Although it may have possibly been grown on the island earlier than the date cited above, it 
may not be a long standing introduction, for, as Christophersen (1968: 4) pointed out, the 
native tussock (Spartina arundinacea) had not been entirely superseded by Phorium tenax as 
a source of thatch when the Norwegian expedition visited the island in 1937-38. Wace (in 
litt.) states that Spartina arundinacea is now never used, and Phormium tenax only 
occasionally, for thatching today (Fig. 21). On Nightingale P. tenax has become ‘the most 
aggressive of the aliens’ now on that island, forming large clumps amongst the tussock grass- 
land and in the Phylica groves and by the late 1960s ‘threatened to spread all over the island, 
thus altering irreversibly the nesting environment of the bulk of the world population of the 
Great Shearwater (Puffinus gravis)’. Steps were therefore taken in 1968 to destroy the flax 
clumps, and this work is being continued by the Tristan islanders (Wace, 1976:60). 
Tristan: Christophersen 1780 (O); Stableford 18 (K). Nightingale: No material seen. 


JUNCACEAE 


Juncus bufonius L. bet 
Introduced. First collected on Tristan da Cunha by Dyer in 1937 but it is so much more 
abundant now, as to have become a common weed of the Settlement Plain down to the 


374 E. W. GROVES 


Fig. 21 Thatching one of the older cottages on Tristan da Cunha with New Zealand flax, Phormium 
tenax. George Glass holds a bundle of flax leaves ready for fixing on the roof. October 1966. 
Photograph: George Edwards. 


beach. It has also been found at Sandy Point, ‘indicating that a halophilous form is 
represented’ (Christophersen, 1968: 3). 

Tristan: Dyer 3524 (PRE); Mejland 127 (O, BM), 137 (O), 1155 (O), 1647 (O, BM, K); Stableford 16 
(K), 51 (BM). 


Juncus effusus L. 
Introduced. Recorded from Tristan da Cunha at Sandy Point, first in 1938 and again in 
1953-54. 

Tristan: Mejland 1134 (O, K), 1136 (O, BM); Stableford 20 (K), 51 (BM), 52 (BM). 


Juncus tenuis Willd. 


J. macer Gray; J. tristanianus Hemsley 
Introduced. First collected by Macgillivray on Tristan da Cunha in 1852. His material was 
described by Hemsley (1885: 154), as J. tristanianus but has since been recognised as being 
the same as J. tenuis. It occurs in wet places at the Settlement, the Potato Patches, and has 
also spread to the upper forest region (Christophersen, 1968: 4). 

Tristan: Christophersen 274 (O, BM, K), 292 (O), 1268 (O); Dickson 101 (BM, AAS); Dyer 3526 
(PRE, NBG); Fleming 81 (E); Keytel 1834 (NBG, K, BM); Macgillivray 343 (K -holotype of J. 
tristanianus), Mejland 116 (O, BM), 127b (O, K), 165 (O), 167b (O), 619 (O, BM, K) 620 (O), 1657 (O, 
BM). 


Rostkovia tristanensis Christoph. 

Endemic. First collected on Tristan da Cunha by Mejland in February 1938 during the 
Norwegian Scientific Expedition 1937-38. The material was subsequently described by 
Christophersen (1944: 3) as a new species, although it is closely related to R. magellanica of 
Patagonia, the Falkland Islands and New Zealand. Specimens found on Gough Island by 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 375 


Wace in 1956 are also thought to be probably this species. On both islands it grows in damp 
upland Empetrum heath; on Gough it is frequent but rarely flowers. 
Tristan: Mejland 1516 (O -holotype, BM, K). Gough: Wace 97 (BM), 132 (BM). 


SCHEUCHZERIACEAE 


Tetroncium magellanicum Willd. 

Native. First found on Gough Island by Wace in December 1955, when it was noted as being 

frequent in the plateau peat of bogs at 370 m and above. No male plants were observed. 
Gough: Wace 22 (BM), 76 (BM), 99 (BM), G.303 (K). 


CYPERACEAE 


Carex insularis Carmich. 
Endemic. First collected on Tristan da Cunha in 1816-17, by Carmichael, who described it 
as new (Carmichael, 1819: 508). It has been found several times on Tristan and also on 
Inaccessible, Nightingale and Gough Islands. On Tristan da Cunha it is frequent amongst the 
tree-fern scrub at about 600 m, and on the other islands it has been found in tussock on 
slopes and in most places. 

Tristan: Bonomi 3 (NBG [ex herb. Mus. Austro-Afric. 1120], K); Carmichael s.n. (K -holotype); 
Christophersen 33 (O, BM), 67 (O, BM, K) 216 (O), 533 (O, BM, K); Dickson 113 (BM, AAS); Mejland 
1151 (O, BM, K); Wace, T.77 (BM), T.239 (K). Inaccessible: Christophersen 2379 (O, BM); Moseley 
s.n. (BM, K, E). Nightingale: Christophersen 2064 (O, BM, K), 2183 (O). Gough: van der Merwe 54 (K); 
Wace 6 (BM), 32 (BM), 127 (BM). 

Carex thouarsii Carmich. 

Endemic. First collected on Tristan da Cunha by Carmichael in 1816-17, and described as 
new. It has since been found on Inaccessible, Nightingale and Gough Islands. It occurs in 
boggy peat mostly above the tree zone, up to 760 m. A variety, var. recurvata, has been 
described by Christophersen (1944: 1), being smaller with recurved leaves. 

var. thouarsii 

Tristan: Carmichael s.n. (K -holotype); Christophersen 66 (O, BM, K), 451 (O, BM, K); Dickson 8 
(BM), 152 (BM, AAS); Dyer 3752 (PRE, NBG); Fleming 75 (E); Macgillivray 337 (K); Mejland 636 
(O); 1171 (O), 1178 (O, BM); Stableford 54 (BM); Wace T.7 (BM), T.16 (BM), T.75 (BM), T.171 (K). 
Inaccessible: Christophersen 2305 (O, BM, K), 2562 (O, BM, K). Nightingale: Christophersen 2035 (O, 
BM), 2225 (O, BM, K); Moseley s.n. (BM, K, E); Wace N.25 (K). Gough: Fleming 7 (E), 13 (E); Wace 
53 (BM), 57 (BM), 85 (BM), 116 (BM), G.239 (K). 
var. recurvata Christoph. 

Tristan: Christophersen 1022 (O -holotype, BM, K); Mejland 1414 (O, BM), 1542 (O, BM, K), 
Gough: Wace 120 (BM). 


Cyperus congestus Vahl OLD BULL GRASS 

Mariscus congestus (Vahl) C. B. Clarke 
Introduced. First collected on Tristan da Cunha by Bonomi in 1904 and recorded from 
several localities on the island since, including the Potato Patches. Hooper (1968: 2) 
considers that it may have been introduced with the advent of potato cultivation. In 1968 
Wace found it colonising cinders and gravel on the side of the new volcano. It occurs also on 
Inaccessible Island. 

Tristan: Bonomi 4 (K, NBG [herb. Mus. Austro-Afric. 1121]); Christophersen 279 (O); Dyer 3542 
(PRE, NGB, K); Keytel 1850 (NBG, K, BM); Mejland 126 (O), 174 (O), 1174 (O); Stableford 53 (BM); 
Wace T.111 (K), T.117 (K), T.122 (K), T.162a (K), T.326 (K). Inaccessible: Christophersen 2483 (O, 
K); Wace 1.17 (K). 


Cyperus esculentus L. NEW BULL GRASS 
Introduced. Collected on Tristan da Cunha by Wace in 1968. It is a very aggressive weed of 
the Potato Patches, where it spreads by root tubers and forms complete cover in neglected 
patches within a few months of the cessation of potato cultivation. 

Tristan: Wace T.257 (K), T.162b (probably this species) (K). 


376 E. W. GROVES 


Cyperus longus L. NEW BULL GRASS 
Probably introduced. Found on Tristan da Cunha in 1968, growing near the Settlement on a 
stream bank by Big Watron. 

Tristan: Wace T.146 (K). 


Cyperus tenellus L_f. 
Introduced. First collected on Tristan da Cunha by Bonomi in February 1904. Like C. 
congestus and C. esculentus, it occurs as a weed in the Potato Patches and was probably 
introduced with cultivation. . 

Tristan: Bonomi 13 (K), 14 (K); Christophersen 221 (O, K), 405 (O, K), 1332 (O); Mejland 131 (O, 
K), 134 (O), 177b (O, K), 178 (O, K), 1422 (O), Stableford 32 (BM). 


Scirpus bicolor (Carmich.) Sprengel* 
var. bicolor 

S. prolifer Thouars, non Rottb.; S. thouarsianus J. A. Schultes; S. thouarsianus var. bicolor 

(Carmich.) Hemsley;?S. thouarsianus var pallescens Hemsley; Isolepis bicolor Carmich.; I. 

squarrosa Carmich.; S. prolifero-ramosus Boeck. 

Endemic. First collected on Tristan da Cunha by Aubert Du Petit-Thouars in 1793. 
Abundant in the plateau mire and wet peaty uplands on Tristan, Inaccessible, Nightingale 
and Gough Islands. It was reported by Dickson in 1966 to be the most successful colonist on 
the new volcanic material (Wace 1967: 51). On Nightingale Island in 1968 Wace found that 
it formed a hummocky meadow, with some individual plants assuming a pachycaul habit 
with ‘trunks’ to c. 50 cm high. 

Tristan: Bonomi 1 (K, PRE, NBG [herb. Mus. Austro-Afric. 1117]), 5 (K, PRE, NBG [herb. Mus. 
Austro-Afric. 1119]); Carmichael s.n. (BM -holotype of I. bicolor), s.n. (BM -holotype of I. squarrosa), 
s.n. (BM); Christophersen 10 (O, K), 59 (O, K), 448 (O, K), 518 (O), 520 (O, K), 524 (O, K), 540 (O, K), 
542 (O, K), 545a (O, K), 546 (O, K), 550c (O, K), 1204 (O, K), 1109 (O, K), 1111 (O, K), 1120 (O), 
1266 (O, K), 1274 (O, K); Dyer 3543 (PRE, NBG, K), 3571 (PRE, NBG, K), 3571 bis (K); Macgillivray 
341 (BM, K), 342 (K); Mejland 110 (O, K), 142 (O, K), 144 (O, K), 603 (O, K), 624 (O, K), 
626 (O, K), 627 (O, K), 638 (O, K), 639 (O, K), 820 (O, K), 1141 (O, K), 1153 (O, K), 1117 (O, K), 1417 
(O), 1424 (O), 1541 (O, K), 1650 (O); 1762 (O, K), 1764 (O, K); Milne s.n. (K); Moseley 6 (BM, K), 8 
(BM, K -both syntypes of S. prolifero-ramosus), s.n. (K); Siggeson 50 & 51 (O)+; Stableford 93 (BM), 
131 sterile (BM); Thouars s.n. (P -holotype of S. prolifer); Wace T.8 (BM), T.9 (BM), T.18 (BM), T.26 
(BM), T.29 (BM), T.102 (K), T.116 (K), T.120 (K), T.226 (K), T.234 (K), T.249 (K), T.252 (K), T.259 
(K). Inaccessible: Christophersen 2327 (O), 2392 (O, K), 2444 (O, K), 2474 (O, K), 2486 (O), 2519 (O, 
K), 2523 (O, K), 2524 (O, K), 2561 (O, K), 2585 (O, K); Moseley 8bis (BM, K -both syntypes of 
S prolifero-ramosus: Wace 1.5 (K). Nightingale: Christophersen 2038 (O, K), 2043 (O, K), 2044 (O, 
K), 2094 (O, K), 2113 (O), 2169 (O, K), 2185 (O, K), 2235 (O, K); Moseley s.n. (K), 4 (K -holotype of S. 
thouarsianus var. pallescens, BM, E). Gough: Brown s.n. (K, E); van der Merwe 22 (PRE, K); Wace 16 
(BM), 108 (BM), 121 (BM), G.216 (K), G.225 (K). 
var. virens (Boeck.) Hemsley 

S. virens Boeck.; S. thouarsianus var. virens (Boeck.) Hemsley; S. oliveri Boeck. 

Endemic variety. First collected on Inaccessible Island by Moseley in 1873, and since found 
on all of the other islands. It occurs in moist situations similar to those of the typical variety 
and differs in its slender habit and by the distinct green apiculus of its glumes. 

Tristan: Bonomi 2 (K, NBG [ex Herb. Mus. Austro-Afric. 1118]); Christophersen 57 (O, K), 1119 
(O); Dickson 42 (BM, AAS), 93 (BM); Dyer 3571 (PRE, NBG, K), 3573 (PRE, NBG, BM, K); Mejland 
1157 (O, K), 1163 (O, K). Inaccessible: Christophersen 2424 (O, K), 2437 (O, K), 2494 (O), 2588 (O, 
K), 2623 (O); Moseley s.n. (K -holotype of S. virens, BM), s.n. (K -holotype of S. oliveri). Nightingale: 
Christophersen 2002 (O, K), 2096 (O, K), 2114 (O, K), 2237 (O, K). Middle: Christophersen 2024 (O, 
K), 2025 (O, K). Stoltenhoff: Christophersen 2154 (O, K), 2155 (O, K), 2156 (O, K). Gough: Brown s.n. 
(K, E); Wace 18 (BM), G.233 (K), G.240 (K). 


_*As has been pointed out by Hooper (1968: 5-6). S. bicolor exhibits a wide variation in size, growth 
form and glume coloration, as well as often occurring in a sterile or proliferous condition—a situation 
which has led to the description of a number of species and varieties all based upon small differences. 
In accordance with Hooper’s treatment only the var. virens Boeck., in addition to that of the type 
variety, is maintained here. 


+ Not seen at O in 1973, but cited as being present by Christophersen (1937: 8-9). 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 397 


Scirpus chlorostachyus Levyns 

Isolepis subtilis Kunth; Scirpus cernuus var. subtilis (Kunth) C. B. Clarke 
Possibly native. First collected on Tristan da Cunha by Macgillivray in November 1852 but 
not noted from any of the other islands in the group. This small, easily overlooked, species 
may be distinguished from both Scirpus cernuus and S. verruculosus (with which it is often in 
association) by its triangulate achenes. On Tristan it has been found in peat in boggy ground 
near Jenny’s Watron, in wet hollows in grassland near the Settlement and at the Potato 
Patches, and in several other damp places. 

Tristan: Christophersen 278 (O, K), 1326 (O, K); Dickson 44 (BM), 81 (BM); Keytel s.n. (NBG*); 
Macgillivray 340 (K), 340b (K), s.n. (BM); Mejland 143a (O, K), 177a (O, K), 605 (O, K), 1142 (O, K), 
1159 (O, K); Milne s.n. (K), Moseley 3 (BM); Siggesson 49 (Ot); Wace T.17 (BM), t.86 (BM), 
T.103 (K). 


Scirpus sulcatus Thouars 
var. sulcatus 

Isolepis sulcata (Thouars) Carmich. 

Endemic variety. First collected on Tristan da Cunha by Aubert Du Petit-Thouars in 1793, 
where it grows in damp and boggy situations or at the margins of pools, often associated with 
Sphagnum. In 1968 Wace collected specimens that were colonising new lava debris. The 
type variety is common on most of the main islands at altitudes from sea-level up to c. 
670 m. 

Tristan: Carmichael s.n. (BM); Christophersen 48 (O, K), 220 (O, K), 525 (O, K), 532 (O, K), 545b 
(O, K), 1023 (O, K), 1095 (O), 1107 (O, K), 1327 (O, K), 1328 (O, K); Dickson 29 (BM), 119 (BM), s.n. 
(BM); Dyer 3574 (PRE, NBG, BM, K); Fleming 71 (E); Keytel 1849 (K); Macgillivray s.n. (BM), 339 
(K); Mejland 347 (O, K), 604 (O, K), 625 (O, K), 1110 (O, K), 1143, (O, K), 1144 (O), 1145 (O, K), 
1146 (O, K), 1154 (O, K), 1169 (O, K), 1549 (O), 1576 (O), 1763 (O, K); Milne s.n. (K); Moseley | 
(BM), s.n. (K); Stableford 19 (K), 19c (K); Thouars s.n. (P -holotype); Wace T.19 (BM), T.38 (BM), 
T.81 (BM), T.100 (K), T.101 (K), T.222 (K), T.248 (K), T.260 (K). Inaccessible: Christophersen 2374 
(O, K), 2396 (O), 2439 (O, K), 2443 (O, K), 2456 (O), 2457 (O), 2471 (O, K), 2499 (O, K), 2586 (O, K); 
Dickson 185 (BM); Moseley 2 (BM), s.n. (K); Wace 1.21 (K). Nightingale: Christophersen 2036 (O, K), 
2099 (O, K), 2100 (O, K), 2101 (O, K), 2102 (O, K), 2133 (O, K), 2134 (O); 2215 (O, K), 2223 (O, K); 
Dickson 136 (BM, AAS); Wace N.2 (K), N.4 (K). Gough: Brown s.n. (K); Discovery ‘W.S.’ Exped. s.n. 
(BM); Wace 17 (BM), 33 (BM), 82 (BM), 88 (BM), G.237 (K), G.247 (K), s.n. (K). 


var. moseleyanus (Boeck.) Hemsley 

S. moseleyanus Boeck. ab 
Endemic variety. This is more slender than the typical variety, having a distinctly acuminate 
glume. Hemsley (1885: 155-6) reports with some reservation (as he had seen only a single 
specimen), that it is ‘easily recognized by its deep red-brown leaf-sheaths’. 


Tristan: Macgillivray s.n. (BM). Inaccessible: Glass in Christophersen 1286 (O, K); Moseley 9 (BM, 
E, K pro parte), s.n. (K), 2 (E). Nightingale: Christophersen 2039 (O, K), 2170 (O, K), 2171 (O, K), 
2184 (O, K), 2222 (O, K), 2234 (O, K); Dickson 143 (BM); Moseley 7 (K -holotype, BM, E), s.n. (K); 
Stableford s.n. (K); Wace N.19b (K). Gough: Brown s.n. (K, E); Discovery ‘W.S.’ Exped. s.n. (BM); van 
der Merwe 22 (PRE—probably this variety); Wace 33a (BM), 88a (BM). 


Scirpus verruculosus Steudel 

Isolepis verruculosa (Steudel) Nees 
Native. First collected on Tristan da Cunha in 1873 by Moseley during the Challenger 
Expedition but not re-found until the visit of the Norwegian Scientific Expedition in 1937. 
However it has probably been overlooked for it frequently occurs in association with Scirpus 
chlorostachyus. 


Tristan: Mejland 130a (O, K), 130b (O), 143b (O, K), 602 (O); Moseley s.n. (K); Wace T.155 
(K). 


*Not among sheets received (1975) on loan from NBG, but cited by Philips (1913: 99) as being collected from 
around the Settlement. ‘ : ¢ 
+Not seen at O in October 1973, but recorded by Christophersen as being there (Christophersen 1937: 8). 


378 E. W. GROVES 


Uncinia brevicaulis Thouars var. brevicaulis 

U. brevicaulis var. gracilior Hemsley; U. brevicaulis var. rigida (Boeck.) Kiikenthal; U. gracilis 

Thouars, nom. illeg. 

Native variety. First collected on Tristan by Aubert Du Petit-Thouars in 1793. Two varieties 
of brevicaulis have been described, the type variety being the taxon represented in the 
Tristan-Gough group and also on the islands of Amsterdam and St Paul, and the var. 
macloviana (Gaudich.) C. B. Clarke occurring on the Falkland Islands, Juan Fernandez and 
in Chile. The two varieties differ mainly in the shape of their bracts. In the former the bract is 
oblong and acute to acuminate and the utricle exceeds the glume in length, whilst in the 
latter the bracts are obtuse and the utricle is seldom longer than the glume. In the 
Tristan-Gough group var. brevicaulis is frequent in wet peat and bogs, and occasionally 
amongst damp grass in a mixed Phylica/Blechnum association. 

Tristan: Carmichael s.n. (K, BM, O); Christophersen 453 (O, BM, K), 459 (O, BM, K), 1030 (O, 
BM). 1065a (O), 1094 (O), 1573 (O); Dickson 25 (BM, AAS); Dyer 3575 (PRE, NBG, K); Macgillivray 
338 (K -holotype of U. brevicaulis var. gracilior), s.n. (BM); Mejland 824 (O, BM), 1179 (O, BM), 1358 
(O), 1372 (O, BM), 1473 (O, BM, K), 1614 (O, BM); Milne s.n. (K); Stableford 2 (K); Thouars s.n. (P 
-holotype of U. brevicaulis); Wace T.37 (BM), T.221 (K). Inaccessible: Christophersen 2304 (O), 2390 
(O, BM, K), 2473 (O, BM), 2563 (O, BM). Nightingale: Christophersen 2218 (O, BM, K); Wace N.32 
(K). Gough: van der Merwe 54a (K); Wace 29 (BM), 58 (BM), 83 (BM), 136 (BM). 


Uncinia compacta var. elongata C. B. Clarke 

Native variety. First collected on Tristan da Cunha by Christophersen in December 1937. 
The typical variety occurs in Australia (New South Wales and Tasmania) whilst var. 
elongata 1s found only on the islands of the Tristan-Gough group and on Amsterdam and St 
Paul Islands. A summary of their differences is given by Hooper (1968: 7). In the Tristan- 
Gough group var. elongata grows in moss on bogs and also on damp cliffs, often in associ- 
ation with mats of Empetrum. 

Tristan: Christophersen 43 (O, K), 64 (O, K), 516 (O, K), 1065 (O, K); Dickson 118 (BM, AAS); 
Mejland 633 (O, K), 1368 (O, K), 1371 (O, K), 1373 (O), 1379 (O, K), 1387 (O, K), 1517 
(O, K), 1557 (O, K), 1614a (O), 1654 (O, K); Wace T.168 (K), T.228 (K). Inaccessible: Christophersen 
2387 (O, K), 2470 (O), 2477 (O, K). Gough: Wace 24 (BM), 95 (BM), 113 (BM), 122 (BM), G232 (K 
-probably this species). 


Uncinia meridensis Steyerm. 

U. smithii Philcox 
Native. First collected on Tristan da Cunha by Mejland in 1937. This small perennial sedge 
with a creeping rhizome and glaucous, crisp leaves, is frequent in bogs and wet heaths, 
particularly between 100-900 m where it often forms tufts or mats amongst Rhacomitrium 
moss and hepatics. In 1968 it was found also on Gough Island. See Hooper (1968: 7-8) for 
affinities; it is possible that Uncinia sinclairei from New Zealand is closely related. 

Tristan: Mejland 1357 (O, K), 1615 (O, K); Wace T.214 (K), T.227 (K), T.229 (K). Gough: Wace 
G.231 (K). 


GRAMINEAE 


Agrostis carmichaelii J. A. Schultes & J. H. Schultes 

Agrostis ramulosa Carmich. (1819), non. A. ramulosa (Kunth) Roemer & J. A. Schultes (1817). 
Endemic. First collected on Tristan da Cunha by Carmichael in 1816 and described as A. 
ramulosa (Carmichael, 1819: 504). The species occurs in damp places. 

Tristan: Carmichael s.n. (BM -holotype of A. ramulosa); Christophersen 47 (O, BM, K), 205, (O, K), 
736 (O, K), 1265 (O, BM, K); Macgillivray 346 (K); Mejland 634 (O, BM, K), 637 (O, BM, K), 822 (O, 
K), 1382 (K), 1385 (K), 1562 (O, BM, K), 1590 (O, K), 1765 (O, K); Wace T.219 (K), T.318 (K 
-probably this species), T.348 (K). Inaccessible: Christophersen 2391 (O, BM, K), 2480 (O, K), 2505 
(O, K). Gough: Wace 142 (BM), G.210 (K), G.249 (K). 


Agrostis castellana var. mixta Hackel 
Introduced. This grass is a loosely tufted perennial, with slender to somewhat stout rhizomes, 
stoloniferous, 30-70 cm high; culms slender to relatively stout; leaf-sheaths smooth; ligules 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 379 


of culm-leaves 1-3 mm long; blades up to 20 cm long and 2.5 mm wide. Panicles loose or 
finally contracted, variable in size, up to 16 cm long. Spikelets 2-3 mm long, those at the tips 
of the branches and branchlets awned, the lateral ones awnless; lemmas of terminate 
spikelets 3—5-nerved with the side-nerves minutely excurrent, sometimes pubescent on the 
sides, awned from just above the base, the awn geniculate, 1.5-3 mm long; basal callus 
minutely bearded; palea } to 3 the length of lemma; lemmas of lateral spikelets glabrous, 
3-nerved, awnless. 

It was first found for the group on Tristan da Cunha by Keytel during 1908-09. Since that 
time it has been collected on that island on a number of occasions, not only in pasture at the 
Settlement but also on the Peak slope, where it was observed as forming a dense sward on 
small ridges leading up a gully. It has also been collected once on Gough Island. 

This grass may have been introduced from the Mediterranean region, or via South Africa, 
or some other part of the world where it has become established. It is often confused with 
Agrostis tenuis. The var. mixta is frequently found among the seeds of the turf-forming grass 
‘Highland Bent’, which are exported from the United States. This and other varieties of A. 
castellana hybridise with A. tenuis yielding a wide range of intermediates. 

Tristan: Christophersen 300a (O), 402a (O); Dyer 3534 pro parte (PRE, BM, K), 3551 (O, K, PRE, 
NBG); Keytel 1838 (K); Mejland 173 (O, BM). Stableford 7 pro parte. Gough: van der Merwe 71 pro 
parte (PRE, K). 


Agrostis crinum-ursi Mez 

Endemic (?). A grass from Tristan da Cunha gathered by an unknown collector (prior to 
1921) and deposited in the Berlin herbarium, was described by Carl Mez (Reprium Spec. nov. 
Regni veg. 17: 300 (1921)) under the above name. In his description Mez gave it as being a 
densely caespitose perennial, up to 17 cm high, with finely filiform-subulate rather rigid 
leaf-blades, up to 4 cm long, few spiculate inflorescences 2.5cm long by 5 mm wide, and 
spikelets up to 2.2 mm long, with awnless 5-nerved lemmas. It may be one of the imperfectly 
known tufted species since gathered at the islands, but unfortunately the type material was 
destroyed in 1943. 


Agrostis gigantea Roth 

Introduced. This species was first collected on Tristan da Cunha by Mejland during the 
Norwegian Scientific Expedition of 1937-38. It was found growing near the Settlement. A. 
gigantea has been gathered again there since and also on Inaccessible Island, near a ruined 
cottage at Saltbeach. 


Tristan: Dickson 97 (K); Mejland 2199 (O, K). Inaccessible: Dickson 130 (K). 


Agrostis goughensis C. E. Hubbard, sp. nov. (Fig. 22) 
a A. stoloniferae spiculis longioribus, lemmatibus pubescentibus aristatis, callis dense barbatis differt. 
Perennial, forming mats. Culms up to 80 cm long, ascending from a creeping base, rooting 
at the nodes, stoloniferous, up to 7-noded, comparatively stout, terete, glabrous, smooth. 
Leaf; sheaths finally shorter than the internodes, glabrous, smooth; ligules oblong, obtuse, 
membranous, up to 10 mm long; blades narrowly linear, tapering to an acute tip, flat, up to 
25 cm long and to 6 mm wide, finely ribbed above, scaberulous on the ribs and margins, 
smooth beneath, green. Panicles oblong, inclined, rather congested, up to 23 cm long and to 
3 cm wide; rhachis scabrid; branches densely clustered, unequal, spreading, up to 5 cm long, 
divided, with filiform scaberulous, closely spiculate branches and branchlets; pedicels 
1-4 mm long, scaberulous. Spikelets 44.5 mm long, lanceolate, acute, narrowly oblong- 
lanceolate when opened out, keeled, scaberulous on the keel, membranous, 1|-nerved, the 
lower as long as the spikelet, the upper narrower than the lower and slightly shorter; lemma 
2-2.3 mm long, ovate-oblong when opened out, obtuse, hyaline, finely 5-nerved, finely and 
appressedly thinly pubescent, awned from the upper third of the lemma; awn straight, 
minutely scaberulous, fine, up to 2.5 mm long; callus densely bearded with white hairs up to 
0.5mm long. Palea ovate-oblong, obtuse, 1 mm long, hyaline finely 2-nerved. Anthers 
0.8-1 mm long. 


380 E. W. GROVES 


Fig. 22 Agrostis goughensis C. E. Hubbard. a. habit (x 4); b. ligule (x 3); c. spikelet (x 10); d. glumes 
(x 10); e. floret (x 10); f. lemma (x 10); g. palea (x 10); h. stamens and ovary (x 13). All drawn 
from Wace 49. 

Fig. 23 Agrostis holgateana C. E. Hubbard. a, habit (x 3); b. ligule (x 7); c. raceme (x 3); d. spikelet 

(x 7); e. glumes (x 8); f. lemma (x 8); g. palea (x 8); h. grain (x 7). Drawn from Wace T. 235. 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 381 


Endemic. First found for the group on Gough Island in December 1955 by Wace, who 
discovered it forming mats with stolons over freshwater pools and around wet places. It has 
the same stoloniferous habit and similar long ligules to that of A. stolonifera L. subsp. 
stolonifera, but its spikelets are larger and have thinly pubescent awned lemmas, and a 
minutely bearded callus. 

Gough: Wace 49 (BM -holotype). 


Agrostis holgateana C. E Hubbard, sp. nov. (Fig. 23) 
a A. mediae Carmich., habitu dense caespitoso, culmis rigidis, laminis setaceis, paniculis linearibus 
vel lanceolato-linearibus densis, spiculis longioribus differt. 

Perennial, densely caespitose, 9-28 cm high; innovations intravaginal. Culms erect or 
slightly spreading, stiff, filiform, branched and several-noded in the lower part, leafy, 
glabrous, smooth, the uppermost internode exserted. Leaf-sheaths overlapping, glabrous, 
smooth, glossy; ligules ovate-oblong to oblong, obtuse or truncate, |-2.5 mm long, firmly 
membranous to chartaceous; blades setaceous, obtuse to subacute, stiff, erect, tightly 
involute, straight or slightly curved, 3-6 cm long, 0.4 - 0.6 mm in diameter, scaberulous on 
the margins. Panicle linear to lanceolate-linear, spike-like, erect, 1.5-3.5 cm long, up to 
4 mm wide, dense; rhachis smooth; branches unequal, erect, 2—3-nate, up to 15 mm long; 
pedicels 0.5-2.5mm long. Spikelets 4-5 mm long, narrowly oblong, sharply acute, 
appressed to the branches; glumes unequal, sharply acute, keeled, prominently scabrid on 
the keel and on the surfaces, narrowly lanceolate in profile, narrowly oblong-lanceolate 
when opened out, coriaceous, the lower as long as the spikelet, with the tip slightly recurved, 
l-nerved, the upper 0.5-1.5 mm shorter than the lower, also narrower, finely 3-nerved. 
Lemma ovate-elliptic when opened out, 1.6-1.7 mm long, minutely 2-lobed, or truncate and 
minutely denticulate, 3-S-nerved, with the middle nerve ending in a mucro or a short 
straight awn up to 0.8 mm long, smooth, membranous: callus minute, glabrous. Palea about 
two-thirds the length of the lemma, oblong, truncate-emarginate, membranous, finely 
2-nerved, smooth. Grain 1-1.2 mm long, lanceolate to oblong in outline. 

Endemic. First collected for the group on Tristan da Cunha in 1968 by Wace who found it 
at three localities: to the north and east of Round Hill (including Soggy Plain), beside a small 
tarn in the crater of Green Hill, and at the base of Tristan Peak at 1060 m. At each of these 
sites the grass was found growing in damp conditions associated with sedges and mosses 
including Sphagnum. Agrostis holgateana is distinguished from other Tristan species of 
Agrostis by its stiff dense form of growth, its very slender rigid culms and leaf blades, and its 
erect spike-like panicles. It has been named in honour of Dr Martin W. Holgate, leader of the 
1955 Gough Island Scientific Survey Expedition. 

Tristan: Wace T.177 (K), T.218 (K), T.206 (K), T.235 (K -holotype). 


Agrostis lachnantha Nees 
Introduced. First collected in 1957 by van der Merwe and then by Wace in 1968. Both 
collectors found it growing by paths and on disturbed ground near the Meteorological 
Station. 

Gough: van der Merwe 62 (PRE, K); Wace G.251 (K). 


Agrostis magellanica subsp. laeviuscula C. E. Hubbard, subsp. nov. 

a subsp. magellanicae, rhachidi et ramis et ramulus et pedicellis laevibus, ghumarum carinis breviter 

scabridis vel fere laevibus, aristis brevioribus strictis, paleis minoribus differt. 
Perennial, stoloniferous or tufted, 10-30 cm high; stolons wiry. Culms tufted or ascending 


from a many-noded trailing base, slender, several-to-many-noded, closely branched or 
simple, closely sheathed, glabrous, smooth. Leaf-sheaths striate, longer than the internodes, 
glabrous, smooth; ligules oblong, very obtuse, thinly membranous, 1|.5-6 mm long, 
becoming lacerate; blades linear, pungent-acute or obtuse, 2-16 cm long, |-4 mm wide, flat, 
firm, green, closely and prominently ribbed above, with the ribs minutely hispid or scabrid, 
smooth and glabrous beneath. Panicles lanceolate to narrowly oblong, contracted and rather 
dense to somewhat lax, or slightly lobed, green, 4-16 cm long, 1.5-3 cm wide; rhachis 
smooth, glabrous; branches fascicled, divided, smooth, glabrous; pedicels thickened towards 


382 E. W. GROVES 


their tips, 24 mm long, smooth. Spikelets up to 5.5 mm long, narrowly lanceolate to 
narrowly oblong, or finally gaping. Glumes very finely acute, shortly aristate-acuminate, 
straight or with the tips slightly recurved, keeled, glabrous, slightly scabrid on the keels, 
otherwise smooth, 1-nerved, firmly membranous except for the thinner margins; lower 
lanceolate opened out, 4.5-5.5 mm long; upper from four-fifths to almost as long as the 
lower, narrowly lanceolate opened out. Lemma ovate-elliptic or elliptic, 1.7—-1.8 mm long, 
slightly and obtusely lobed or truncate, thinly membranous, very minutely and obscurely 
verrucose, glabrous, finely 5-nerved, with the lateral nerves sometimes minutely excurrent, 
awned from the back towards the tip, with the awn fine, straight, up to 2.5 mm long, 
minutely scaberulous; palea 0.8-1 mm long, oblong, obtuse; rhachilla not produced; callus 
glabrous. Anthers 0.8 —1 mm long. Grain, oblong, dorsally compressed, |.5 mm long. 

Agrostis magellanica Lam. (Tabl. Encycl. 1: 160 (1791)) was based on plants collected by 
Commerson in the eastern Magellan; it is widespread in the south Antarctic, and differs in 
the ramifications of the panicle being minutely hispidulous, the equal or subequal similar 
glumes scabridly-hispidulous on the keels, and the awns geniculate. 

Endemic subspecies. It was first found both on Tristan da Cunha and on Inaccessible 
Island early in 1938, during the Norwegian Scientific Expedition visit to the group. On 
Tristan it occurs in grassy areas just above the tree-fern limit. It has more recently been 
collected on Gough Island. 

Tristan: Christophersen 1584 (K); Mejland 1364 (O, K, BM), 1551 (K -holotype, O, BM), 1564 (K); 
Wace T.166 (K), T.173 (K), T.173 (K), T.233 (K). Inaccessible: Christophersen 2521 (O, K, BM). 
Gough: Wace G.236 (K), G.246 (K). 


Agrostis media Carmich. 

Endemic. First collected on Tristan da Cunha by Carmichael in 1816-17, who described his 
material as a new species. It occurs frequently over the island on grassy slopes amongst the 
tree-fern scrub between 600 and 800 m, and has also been found on Inaccessible and Gough 
Islands. 

Tristan: Carmichael s.n. (BM -holotype); Christophersen 50 (O, K, BM), 61 (O, K), 211 (O, K), 523 
(O, K, BM), 549b (O, K), 550f (O, K, BM), 550g (O, K, BM), 1325 (O, K); Dickson 33 (K); Mejland 
1366 (O, K, BM), 1367 (O, K), 1381 (O, K, BM), 1382 (O, K, BM), 1384 (O, K, BM), 1385 (O, K, BM), 
1574 (O, K); Milne s.n. (K); Moseley s.n. (K); Wace T.196 (K), T.225 (K), T.318 (K). Inaccessible: 
Christophersen 2472 (O, K), 2508 (O, K). Gough: Brown s.n. (K); Fleming 25 (E); van der Merwe 67b 
(PRE); Wace 72 (BM), 74 (BM), 91 (BM), 123 (BM), G211 (K), G.248 (K), G.249 (K). 


Agrostis stolonifera L. subsp. stolonifera 

A prostrata J. D. Hook.; A. difficilis Hemsley 
Introduced. First collected on Tristan da Cunha by Mejland in January 1938. It has now 
become a common weed on the Settlement Plain and at the Potato Patches, where it is quick 
to spread on disturbed ground. It has also been found on both Inaccessible and Gough 
Islands, where on the former it is now frequent at the foot of the cliffs and around the huts at 
Waterfall Beach, and on the latter it has been found forming mats among native grasses and 
sedges. 

Tristan: Christophersen 296 (O, K, BM); Dickson 91 (K); Mejland 1166 (O, BM, K), 1309 (O, BM, 
K); Stableford 36 (sterile) (BM), 49 (BM); Wace T.80 (BM), T.255a(2) (K). Inaccessible: Wace 1.23 (K), 
1.50 (K); Gough: van der Merwe 71 pro parte (K); Wace 48 (BM), 130 (BM), 134 (BM), G.250 (K). 


Agrostis tenuis Sibth. 

A. simulans Hemsley 
Introduced. First collected on Gough Island when the RRS Discovery II called there in June 
1927 on her homeward voyage after a season of Antarctic surveys. It has also been found on 
several occasions on Tristan da Cunha occurring as a common grass in pastures near the 
Settlement and at the Potato Patches. More recently it was discovered forming a deep sward 
on slopes below the cliffs behind the new volcano. 

Tristan: Christophersen 270a (O), 271 (O), 300 (O), 402 (O), 1131 (O, BM, K), 1331 (O, BM); 
Dickson 22 (BM), 53 (BM), 90 (BM), 173 (BM—probably this species); Dyer 3534 pro parte (K, BM); 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 383 


Mejland 173 (O, BM); Stableford 6 (K); 7 pro parte (K), 48 (BM), 109 (BM); Wace T.108 (K), T.123, 
(K), T.130 (K), T.255a (1) (K). Gough: Discovery ‘W.S’ Exped., s.n. (BM); van der Merwe 71 pro parte 
(PRE); Wace 47 (BM). 


Agrostis trachychlaena C. E. Hubbard, sp. nov. (Fig. 24) 
a. A. stoloniferae L., foliorum lamina supra scabra vel minute hispida et prominenter et arcte striata, 
glumis hispido-scabris demum deciduis differt. 

Perennial, loosely tufted. Culms trailing, probably stoloniferous, ascending from a creeping 
base, 20-55 cm long slender, many-noded, profusely branched, glabrous, smooth. Leaf- 
sheaths overlapping, terete, persistent, prominently striate, the ribs minutely scaberulous or 
smooth, glabrous; ligules oblong, 1-3 mm long, thinly membranous, becoming lacerate; 
blades linear, finely acute, 2-7 cm long, 0.5-2 mm wide, flat, spreading, prominently and 
closely ribbed on the upper surface, with the ribs scabrid or minutely hispid, glabrous and 
smooth beneath. Panicles shortly exserted from the uppermost sheaths, narrowly lanceolate 
to narrowly oblong, contracted and rather dense, 1.5—5.5 cm long, 6-16 mm wide; rhachis 
smooth or slightly hispid; branches 3-S-nate, slightly flexuous, stiff, erect or spreading 
obliquely, divided, 0.3-1.5cm long, scabro-hispid; pedicles thickened upwards, club- 
shaped, 0.3-2 mm long, scabro-hispid, sometimes apparently articulated at the base. 
Spikelets linear to linear-lanceolate, acute, 3-4 mm long. Glumes acute, conspicuously 
hispidly scabrid, firmly membranous, finally deciduous; lower as long as the spikelet, 
lanceolate opened out, 1—3-nerved; upper three-fourths to four-fifths the length of the lower, 
lanceolate opened out, 1-3-nerved. Floret slightly dorsally compressed; rhachilla not 
produced; lemma 1.3-1.8 mm long, ovate-oblong, broadly very obtuse or dentate-truncate, 
thinly membranous, 5-nerved, with the nerves reaching the tip or minutely excurrent; palea 
broadly oblong, very obtuse, minutely scaberulous, 2-nerved, smooth, 0.6-0.8 mm long. 
Grain immature, triquetrous, narrowly oblong, | mm long. 

Endemic. First discovered for the group by Christophersen on both Inaccessible and 
Nightingale Islands, in early 1938, during the Norwegian Scientific Expedition. Found on 
damp rocks or on rocky slopes in the shade of tussock (Spartina arundinacea) where 
sheltered and in damp conditions. Not yet recorded from any of the other islands. It is a most 
distinct species, unlike any other Agrostis from South Africa, tropical Africa, the Mascarenes 
or Atlantic Islands. 

Inaccessible: Christophersen 2578 (K -holotype, O). Nightingale: Christophersen 2213 (O, BM, K); 
Wace N.6 (K), N.33 (K), N.51 (K). 


[Agrostis tropica Beauv. 

Palisot de Beauvois (1813, II, 37 T.80, fig, 2) described this species from a specimen collected by Aubert 
Du Petit-Thouars from Ile de Prince (Principé) and said that the same collector also found it in 
Mauritius. Not only is the type of the grass no longer available for study (although Beauvois’ 
illustration is good) but subsequent collectors have failed to rediscover it on either island. This is 
perhaps understandable as neither has a suitable habitat. Aubert Du Petit-Thouars was, however, also 
known to have visited the islands of Tristan da Cunha and réunion (see p. 347) during the same voyage 
and that some of his specimens, labelled subsequent to collection, became wrongly localised (see note 
under Amauropelta bergiana, p. 404). It is possible, therefore, that the specimen on which Beauvois’ 
Agrostis tropica was based came not from Principé, but from Tristan da Cunha or Réunion. No 
specimen of any Agrostis from Tristan da Cunha seen so far matches Beauvois’ illustration of A. tropica 
(=pubescent glumes, cf. scabrid or glabrous glumes of all the Tristan da Cunha Agrostis), and while the 
possibility of finding it on that island cannot entirely be ruled out, it would seem that Réunion is the 
more likely locality for this grass.] 


Agrostis wacei C. E. Hubbard, sp. nov. (Fig. 25) 

gramen perenne compacte caespitosum glaucum; culmi tenuiter filiformes, flexuosi, multinodes, 
ramosi, laeves; laminae numerosae, tenuiter filiformes, obtusae, breves, teretes, involutae, glabrae; 
ligulae brevissimae; racemi 1|—4-spiculati, breves; spiculae oblongae, 1.6—2 mm longae; glumae 
aequales vel paullo inaequales, minute scabridae; lemma truncatum, minute 5-dentatum, glabrum, 
5-nerve; palea fere lemmate aequilonga. 


384 E. W. GROVES 


s\ 
I 

\ IAG 

“\ Wy 


A VW 
‘i ye 


\ SYA 
aS 
WHY Y 


7 
Me 
WY 
f/ 


CAVA Wh) 
NENA 
STNG 


Fig. 24 Agrostis trachychlaena C. E. Hubbard. a. habit (x 3); b. ligule (x 4); c. spikelet (x 10); d. 
glumes (x 13); e. floret (x 13); f. lemma (x 13); g. palea (x 13); h. stamens (x 13). The habit and 
ligule are drawn from Christophersen 2578; the remainder are drawn from Wace N. 51. 

Fig. 25 Agrostic wacei C. E. Hubbard. a. habit (x 3); b. ligule (x 13); c. spikelet (x 13); d. glumes 
(x 20); e. floret (x 27); f. lemma (x 20); g. palea (x 20); h. stamens and ovary (x 15). All drawn 
from Wace T. 319. 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 385 


Densely caespitose, forming soft hammocks up to 20 cm or more in diameter, up to 12 cm 
high. Culms finely filiform, terete, flexuous, many-noded, branched, leafy, glabrous, smooth. 
Leaves very numerous, glaucous when fresh (ex Wace); sheaths overlapping, longer than the 
internodes, finely ribbed, thin, glabrous, very minutely scaberulous; ligules ovate-oblong or 
oblong, up to | mm long, membranous; blades finely filiform, obtuse spreading obliquely, 
10-14 mm long, terete, involute, glabrous, very minutely and obscurely appressed scaberu- 
lous, rather stiff, curved or straight. Inflorescence a raceme, 1-4-spiculate, 3-7 cm long, 
shorter than the terminal leaves; rhachis straight or slightly flexuous; pedicels up to 1.5 mm 
long. Spikelets oblong or gaping, 1.6-2 mm long. Glumes equal or unequal with the lower 
longer and as long as the spikelet, oblong to lanceolate-oblong, obtuse or acute, keeled, 
minutely scabrid all over, rigid on the keel, otherwise membranous, 3-nerved; lemma 
1.2-1.4 mm long, broadly elliptic, truncate, minutely 5-dentate, awnless, thinly to firmly 
membranous, glabrous, smooth, 5-nerved; cellus very obtuse, glabrous; palea oblong, | mm 
long, glabrous, hyaline; anthers | mm long. 

Named in honour of Dr Nigel M. Wace, botanist on the 1955 Gough Island Scientific 
Survey Expedition. His collections have added measurably to our knowledge of the flora of 
the group. 

Endemic. First discovered for the group on Tristan da Cunha in November 1976 by Wace 
who found it growing amongst clumps of Blechnum palmiforme on the precipices under Big 
Green Hill at c.500 m alt., and forming soft hummocks up to 20 cm or more in diameter and 
very glaucous when fresh. Agrostis wacei is a very distinct grass, apparently not closely 
related to any of the other Tristan or African species of Agrostis. 

Tristan: Wace T.319 (K -holotype, BM). 

Three other examples of Agrostis collected by Wace (T.195, T.223 and T.224) also from Tristan da 
Cunha, represent a species allied to A. wacei but distinguished by their slightly larger, smooth glumes 
up to 3 mm long (cf. glumes in A. wacei up to 2 mm long). The material collected under all the above 
three numbers, however, is unfortunately either sterile or inadequate for description. Further collection 
of this Agrostis is therefore desirable. 


Agrostis sp. 

Phalaris cespitosa Thouars (1808: 37) 
Endemic (?). Amongst the grasses found by Aubert Du Petit-Thouars on Tristan in 1793 was 
one which he regarded as new and gave the following short description (Aubert Du 
Petit-Thouars, 1808: 37, 84), ‘Panicula pauciflora subspicata culmo repente stolonifero. 
Herb, I'®. Petite espéce a chaume rampant stolonifére; panicule pauciflore, resseré en un €pi 
long d’un pouce. Cette plants couvre quelque fois, a elle seule, un grand espace de terrain.’ 
Judging from this brief account, the grass is probably a species of Agrostis with stolons and 
spreading culms and with few flowered, spike-like panicles 2.5 cm long. 

Tristan: Thouars s.n. (P*). 


Aira caryophyllea L. 
Introduced. First collected on Tristan da Cunha in 1937 and on Inaccessible Island in 1938, 
both during the visit of the Norwegian Scientific Expedition. On Tristan da Cunha it has 
been found in pasture grassland around the Settlement, where it has become frequent, and 
also along track sides or as a coloniser of bare patches. On Inaccessible Island it has 
been found growing on the steep slopes above Blenden Hall. 

Tristan: Christophersen 2 (O, K), 109 (O); Dickson 8 bis (K), 16 (BM); Mejland 109 (O); Stableford 
88 (BM), 117 (BM); Wace T.87 (BM), T.105 (K), T.132 (K), T.231 (K), T.311 (K), T.312 (depauperate) 
(K). Inaccessible: Christophersen 2481 (O, BM), 2487a (O). 


Anthoxanthum odoratum L. 
Introduced. First collected on Tristan da Cunha by Siggeson in November 1934. It occurs 
quite commonly in pasture at the Settlement and at the Potato Patches. 

Tristan: Christophersen 272 (O, BM, K); Mejland 161 (O); Siggeson 55 (O); Stableford 12 (K), 12a 
(K), 37 (BM), 86 (BM); Wace T.48 (BM). 


* It was not possible to find the holotype on my visit there in September 1975. 


386 E. W. GROVES 


Bromus willdenowii Kunth 

B. catharticus auct., non Vahl; B. unioloides auct., non (Willd.) Kunth 
Introduced. First collected on Tristan da Cunha in 1852 by Macgillivray and also by Milne, 
both on the voyage of HMS Herald. It occurs on that island by the edges of paths and on 
ungrazed pasture near the Settlement. More recently, in 1968, it has been found growing in 
sand against the new lava blocks at the level of the old sea beach. 

Tristan: Dickson 50 (BM); Macgillivray 347 (K), s.n. (K); Milne s.n. (K); Stableford 2 (K), 41 (BM); 
Wace T.49 (BM), T.152 (K), T.261 (K). 

Calamagrostis deschampsiiformis C. E. Hubbard, sp. nov. (Fig. 26) 

gramen perenne dense caespitosum, 35-50 cm altum; culmi erecti, simplices, 2-nodes; foliorum 
vaginae laeves, basales persistentes, coriaceae; ligulae acutae, 3-7 mm longae; laminae junciformes, 
obtusae, cylindricae, involutae, rigidae, 8-19 cm longae, 1-1.2 mm latae, glabrae; paniculae oblongae 
13 cm longae; rami fasciculatae; pedicelli 14 mm longi; spiculae 8-10 mm longae; glumae explanatae 
anguste oblongae vel oblongae, acutae; inferior l-nervis, superior 3—5-nervis; lemma 4.5-—5 mm 
longum, bilobum, lobis plerumque acute bidentatis, aristatum, chartaceum, 7-nerve, laeve; ariste 
leviter geniculata vel flexuosa, 5.5—-8 mm longa; callus pilis |-1.6 mm longis barbatus; palaea oblonga, 
4.5—5 mm longa, bicarinata, carinis scaberulis; caryopsis 2 mm longa. 

A densely caespitose leafy perennial, 35-50cm high, with numerous intravaginal 
innovations. Culms erect, stiff, moderately slender, 2-noded, simple, glabrous, smooth. Leaf 
sheaths glabrous, smooth; basal sheaths persistent, coriaceous, up to 8 cm long; ligules 
narrowly ovate, acute, 3-7 mm long, firmly membranous; blades junciform, erect, rigid, with 
coriaceous obtuse tips, cylindrical, involute, 8-19 cm long, 1-1.2 mm in diameter, smooth, 
closely ribbed and glabrous above. Panicles oblong, 13 cm long, 2-3 cm wide, loose; rhachis 
smooth below, sparsely scabrid above; branches clustered, up to 9 cm long, more or less 
scabrid; pedicels 1-4 mm long. Spikelets oblong or gaping, 8-10 mm long; rhachilla 
produced as a plumrose bristle up to 3 mm long, sometimes bearing an awn up to 3 mm long 
on a minute vestige. Glumes, slightly unequal, the upper as long as the spikelet, chartaceous- 
membranous, glabrous, smooth or the lower scabrid on the keel; lower linear-lanceolate in 
profile, acute, narrowly oblong when opened out, 7-9 mm long, keeled, 1-nerved; upper 
oblong when opened out, acute, 3-S-nerved. Lemma oblong when opened out, 4.5-5 mm 
long, 2-lobed with the lobes acutely and unequally 2-toothed, or acutely 5-toothed at the 
apex, firmly membranous-chartaceous to stiffly chartaceous, rounded on the back, 7-nerved, 
smooth, awned on the back in the lowest third; callus truncate, bearded with hairs |-1.6 mm 
long, awn slightly geniculate or flexuous, 5.5-8 mm long, scaberulous; palea oblong, 
2-toothed, as long as the lemma, 2-keeled with the keels scaberulous, firmly membranous. 
Caryopsis 2 mm long, oblong in side view, rounded on both sides, dark brown; embryo 
broadly elliptic; hilum oblong. 

Endemic. First found by Wace on Tristan da Cunha in April 1968 and then on Gough 
Island in the following month. On Tristan it grew in the wet peaty margin amongst mosses 
beside Crater Lake on Stony Beach Hill, and on Gough in wet heath vegetation in Gonydale 
beside streams and wet gullies. 

Tristan: Wace T.251 pro parte (K). Gough: Wace G.219 (K -holotype). 

The species of Calamagrostis (sensu lato) are much in need of critical investigation. They 
appear to have had several different lines of evolution, some having Deschampsia, Trisetum, 
or other ancestral genera. The above new species belongs to a group well represented in parts 
of the Southern Hemisphere (South America, New Zealand and southern Australia), some- 
times separated under Deyeuxia, in which the lemma is finally indurated and chartaceous to 
coriaceous. Until the complex has been studied in greater detail, it seems best to treat them in 
the broad sense, e.g. under Calamagrostis. This new species may have originated from a 
species of Deschampsia, such as D. christophersonii, in which a mixture of 2- and |-flowered 
spikelets may be found, sometimes in the same inflorescence. 


Cynodon dactylon (L.) Pers. 


Introduced. First collected on Tristan da Cunha by Dyer in February 1937 and since found 
several times on the island in pasture by the Settlement and above high water on the beach. 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 387 


Wy 
ae; 
i 
I 


i 


Fig. 26 Calamagrostis deschampsiiformis C. E. Hubbard. a. habit (x ); b. ligule (x 4); c. spikelet 
(x 4); d. glumes (x 5); e. floret (x 5); f. lemma (x 5); g. palea (x 5); h. grain (x 8). The habit and 
ligule are drawn from Wace T. 251; the remainder are drawn from Wace G. 219. 

Fig. 27 Deschampsia christophersenii C. E. Hubbard. a. habit (x 3); b. ligule (x 3); c. spikelet (x 4); 
d. glumes (x 4); e. florets (x 4); f. lemma (x 4); g. palea (x 4); h. stamens and ovary (x 8). All drawn 
from Christophersen 551. 


388 E. W. GROVES 


More recently a single patch was found near the huts at Waterfall Beach on Inaccessible 
Island by Wace in February 1968. 

Tristan: Christophersen 15 (O); Dickson 4 (BM), 146 (BM, K); Dyer 3540 (PRE, K, BM, NBG, O); 
Mejland (O, BM, K), 1297 (O, BM); Stableford 9 (K), 40 (BM), 91 (BM), 92 (BM), 114 (BM); Wace 
T.190 (K), T.327 (K), T.342 (K). Inaccessible: Wace I.16 (K). 

Cynosurus cristatus L. 
Introduced. Collected on Tristan da Cunha during 1953-54. It was probably originally 
imported in a mixed pasture seed. 

Tristan: Stableford 15 (K), 34 (BM). 


Dactylis glomerata L. 

Introduced. First collected on Tristan da Cunha in 1954 by Stableford, who considered it to 
have been introduced in 1953 from South Africa. Wace in December 1955 found two 
clumps growing on Gough Island, above Glen Beach. He too considered it to be 
‘apparently recently introduced’. 


Tristan: Stableford 8 (K), 44 (BM); Wace T. 187 (K). Gough: van der Merwe 72 (K); Wace 
45 (BM). 


Deschampsia christophersenii C. E. Hubbard, sp. nov. (Fig. 27) 


a D. mildbraedii Pilger, ligulis brevioribus, foliorum laminis plerumque laevibus, paniculis 
angustioribus anguste oblongis vel lanceolato-oblongis, lemmatibus tenuioribus, callorum 
pilis longioribus differt. 


Densely tufted perennial, 15-40 cm high, with many intravaginal innovations. Culms 
erect, simple, moderately slender, stiff, 2-noded closely sheathed. Leaf-sheaths overlapping, 
glabrous, smooth, the lower coriaceous, persistent, up to 7 cm long; ligules lanceolate, acute, 
2-5 mm long, firmly membranous; blades erect, rigid, setaceously junciform, with the hard 
pungent or obtuse tip, convolute or complicate-involute, 5-20 cm long, 0.8-1.2 mm wide, 
smooth beneath, closely and prominently ribbed above, with the ribs glabrous and smooth, 
2-2.5 mm wide opened out. Panicles dense to rather loose, erect, narrowly oblong to 
lanceolate-oblong, 7-15 cm long, 1-2 cm wide, greenish-yellow; rhachis glabrous towards 
the base, otherwise minutely hispid; branches slender, erect or ascending, up to 6 cm long, 
minutely hispid or glabrescent; pedicels 0.5-4 mm long, glabrous or slightly hispid. Spikelets 
narrowly oblong or gaping, 2-flowered; rhachilla ciliate, continued as a plumose bristle 
beyond second floret and sometimes tipped with a vestige. Glumes similar, equal or nearly so 
or lower slightly longer than upper, chartaceous except for the membranous margins, 
scaberulous on keel above the middle; lower 7.5—10 mm long, narrowly oblong-lanceolate to 
narrowly elliptic opened out, acute, 1-3-nerved; upper 7.5—9 mm long, narrowly elliptic or 
oblong-elliptic opened out, acute, 3-nerved, with the lateral nerves extending to the middle. 
Lower lemma ovate-oblong, 2-lobed, with the lobes unequally and acutely 2-toothed, 
4.5-6 mm long, firmly membranous or coriaceous and smooth below, thinly to firmly 
membraneous and smooth or minutely scaberulous above the middle, 5—7-nerved, awned; 
awn slightly geniculate, 6-8 mm long, minutely scaberulous, arising in the lowest third of the 
lemma; callus densely bearded with white hairs 2—2.5 mm long; palea nearly as long as the 
lemma, narrowly oblong, acutely 2-toothed, minutely ciliate on the keels, thinly 
membranous; anthers |-1.5 mm long. Upper lemma 4—5 mm long, ovate-oblong unequally 
acutely 4-toothed, 5-7-nerved, membranous to coriaceous, otherwise similar to the lower; 
awn 4-6.5 mm long; palea similar to that of lower floret; anthers |-1.7 mm long. 

Endemic. First found for the group on Tristan da Cunha and Inaccessible Island early in 
1938, during the visit to the islands by the Norwegian Scientific Expedition, the several 
collections being made mostly in damp meadows or bogs between 600-900 m. It was again 
found on Tristan da Cunha in February 1962 by Dickson just above Hottentot Gulch and 
not far from the locality of Mejland 1383 (see below). This grass has been named in honour 
of Dr Erling Christophersen, leader and botanist on the Norwegian Scientific Expedition 
1937-38. The botanical and other scientific results of this expedition made an outstanding 
contribution to our knowledge of the Tristan da Cunha group of islands. 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 389 


Tristan: Christophersen 551 (O, K -holotype), 1105 (O, K); Dickson 120a (K); Mejland 1383 (O, K), 
1556 (O, K); Wace T.76 (possibly this species but material sterile) (BM). Inaccessible: Christophersen 
2469 (O, K). 

In addition to the holotype of Deschampsia christophersenii cited above, in which the 
spikelets so far as they have been examined appear to be all 2-flowered, there are other speci- 
mens which closely resemble the holotype collection in most respects, but have 1- or 
2-flowered spikelets in the same inflorescence, while yet others appear to be strietly 
1-flowered. The last mentioned agree in this respect with some species of the genus 
Calamagrostis (Deyeuxia). The genus Deschampsia Beauv. is widely distributed in 
temperate and cold regions of both hemispheres, but occurs only at high altitudes in the 
tropics. Its spikelets are usually 2-, rarely 3- or 4-flowered, with the rhachilla continued 
beyond the uppermost floret as a fine, usually hairy bristle; its lemmas are 2-lobed, 4-toothed 
or lacerate at the apex, 4—7-nerved, awned from or above the base and with the awn 
geniculate or straight, and mostly with a bearded callus at the base of the floret. The 
specimens of Mejland 1556 are in many respects typical of Deschampsia; but although most 
spikelets are 2-flowered, in some the second floret has been found to be much reduced in size 
and moreover sterile, while 1- or 2-flowered spikelets occur in the same panicle. A similar 
condition has been noted in specimens of Christophersen 1105, the second floret being 
variously reduced, smaller and sterile, sometimes represented by only a minute hyaline 
rudiment or at times even without a vestige of the lemma, so that 1- or 2-flowered spikelets 
may be found in the same inflorescence. Much the same mixture has been observed in 
specimens of Christophersen 2469 from Inaccessible Island (plateau near west end, 450 m, 
23/2/1938), although in this gathering 2-flowered spikelets appear to be of rare occurrence, 
while in those with | floret there is a minute rudiment of the second floret. Similarly, in 
specimens of Mejland 1383 from Tristan da Cunha (east of Hottentot Gulch, 600-700 m, 
28/1/1938), 1-or 2-flowered spikelets occur in the same panicle but with the second floret 
imperfect and ranging from a microscopic vestige at the tip of the prolongation of the 
rhachilla to a minute lemma and no palea, or to a larger lemma and palea but with no flower, 
or sometimes with the vestige of a flower. These intermediate plants require investigation 
experimentally when plants become available for cultivation. 


Deschampsia mejlandii C. E. Hubbard, sp. nov. (Fig. 28) 
a D. christophersenii C. E. Hubbard, ligulis multo longioribus, foliorum laminis planis vel convolutis 
latioribus costis minute hispidis; lemmatibus paulo brevioribus minute scaberulis differt. 

Densely tufted perennial, 12-48 cm high. Culms erect, relatively stout, stiff, simple, 
1-2-noded, closely sheathed, glabrous, smooth. Leaf-sheaths persistent at the base, over- 
lapping, coriaceous, glabrous, smooth, broad, up to 10 cm long; ligules lanceolate, 7-15 cm 
long, chartaceous-membranous, becoming lacerate; blades erect, stiff, linear, mostly 
pungent-acute, or subobtuse, with a hard involute coriaceous tip, 5-22 cm long, 2.5-6 mm 
wide, flat or convolute, glaucous-green or green, smooth beneath, prominently and closely 
ribbed above, with the ribs minutely hispid, scaberulous upwards on the margins. Panicles 
dense, erect or nodding, oblong or lanceolate-oblong, soft, silvery or very pale yellow 
(pallid), plumose, 10-25 cm long, 2—3.5 cm wide; rhachis mostly minutely hispid above; 
branches clustered, minutely hispid or rarely glabrous, up to 8 cm long, divided, closely 
spiculate; pedicels minutely hispid, 1-3.5 mm long. Spikelets oblong or gaping, 8-10 mm 
long, 2-flowered; rhachilla produced as a fine plumose bristle up to 1.5 mm long. Glumes 
narrowly lanceolate in profile, similar, equal or nearly so, firm below the middle, 
membranous above and at the margins, 3-nerved, with the lateral nerves below the middle; 
lower 6.5-8.5 mm long, narrowly oblong-lanceolate to narrowly elliptic-lanceolate opened 
out, acutely acuminate, keeled, scabrid above on the keel; upper up to 10 mm long, narrowly 
oblong-lanceolate opened out, acutely acuminate, keeled, scabrid above on the keel. Lower 
lemma 4-5 mm long, ovate to oblong-ovate opened out, unequally and acutely 4-toothed at 
the tip or 2-lobed with each lobe 2-toothed, thinly membranous, finely 5—7-nerved below the 
middle, minutely scaberulous, densely silky white villous on the basal callus with hairs 


390 E. W. GROVES 


Ss 


Fig. 28 Deschampsia mejlandii C. E. Hubbard. a. habit (x 4); b. ligule (x 14); c. spikelet (x 3); d. 
glumes (x 4); e. floret (x 4); f. lemma (x 4); g. palea (x 4); h. stamens and ovary (x 5). All drawn 
from Christophersen 1106. 

Fig. 29 Deschampsia robusta C. E. Hubbard. a. habit (x 4); b. ligule (x 2); c. spikelet (x 4); d. glumes 
(x 5); e. floret (x 5); f. lemma (x 5); g. palea (x 5); h. stamens and ovary (x 10). The habit is a 
composite drawing based on Wace G. 245 and Wace G. 217; the ligule is drawn from Wace G. 
217; the remainder are drawn from Wace G. 245. 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 391 


2-3.5 mm long, awned; awn straight or slightly bent, 6-6.5 mm long, scaberulous, arising 
towards the base of the lemma; palea shorter than the lemma, narrowly oblong, acutely 
2-toothed, 3-4 mm long, 2-keeled, minutely ciliate on the keels, thinly membranous; anthers 
1-1.5 mm long. Upper lemma 3-4.5 mm long, ovate opened out, 2-lobed with the lobes 
unequally and acutely 4-toothed, finely 5—7-nerved below, minutely scaberulous, mem- 
branous, awned; awn straight, 4-6.5 mm long, arising near base of lemma; palea narrowly 
oblong, acute, 2-toothed, 3-3.5 mm long, minutely ciliate on the 2 keels, thinly mem- 
branous; anthers | mm long. 

This grass has been named in honour of Mr Ingvar Mejland, botanical assistant on the 
Norwegian Scientific Expedition, 1937-38. 

Endemic. First found on both Tristan da Cunha and Inaccessible Island by Christophersen 
and Mejland early in 1938 during the Norwegian Scientific Expedition. It occurred at several 
localities on Tristan da Cunha either at the upper limit of, or just above, the tree-fern zone. 
On Inaccessible Island it occurs in a meadow on the plateau near the western end. 

Tristan: Christophersen 531 (O, K), 533b (O, K), 549a (O, K), 826* (O, K), 1106 (O, K -holotype); 
Mejland 1363 (O, K); Stableford s.n. (K); Wace T.15 (BM), T.74 (BM), T.165 (K), T.170 (K), T.176 
pro parte (K), T.220 (K), T.247 (K), T.346 (K). Inaccessible: Christophersen 2764 (O, K). 

Two other Deschampsia gatherings were made by Wace on Tristan da Cunha in October 
1955: T.14 (BM) and T.78 (BM), both found in Rumex/Holcus grassland above Nellie’s 
Hump at 750-1100 m. The first (T.14) might be a depauperate form of D. mejlandi, but its 
leaves seem different, being very short and rolled; the second (T.78) has some of the stems 
rooting at the nodes. Regrettably these gatherings lack spikelets and therefore it is not 
possible to say if they represent a distinct species. Collection of further material of both 
grasses from this locality is therefore desirable. 


Deschampsia robusta C. E. Hubbard, sp. nov. (Fig. 29) 
a D. mejlandii C. E. Hubbard, culmis altioribus robustioribus 2-5-nodis, laminis plerumque 
longioribus junciformibus involutis, paniculis laxiusculis latioribus, ramis scabridis vel laevibis differt. 

A densely caespitose perennial, 80-110 cm high, from a short rhizome. Culms stout, 
terete, up to 5 mm in diameter, simple, 2—5-noded, glabrous, smooth. Leaf-sheaths glabrous, 
smooth; basal sheaths persistent, coriaceous glossy, up to 12cm or more long; ligules 
lanceolate to ovate, acute, 8-15 cm long, membranous; blades junciform, with an acute or 
obtuse coriaceous tip, 20-45 cm long, tightly involute or opening out and up to 7 m wide, 
rigid, closely and prominently ribbed above, with the ribs shortly hispid, glabrous and 
smooth beneath. Panicles oblong-lanceolate to ovate, 15-25 cm long, 4—5 cm wide, inclined, 
somewhat loose: rhachis smooth in lower part, sparsely scabrid above or throughout; 
branches clustered, up to 10 cm long, scabrid or smooth; pedicels 1-6 mm long. Spikelets 
oblong or slightly gaping, 8-11 mm long, 2-flowered; rhachilla plumose, produced as a 
bristle beyond the second floret. Glumes as long as the spikelet, equal or almost so, narrowly 
lanceolate in profile, finely acute, firmly membranous, keeled, lower 1-nerved; upper 
3-nerved, slightly wider than the lower. Lower lemma ovate-oblong when opened out, 
2-lobed, with the lobes unequally and finely 2-toothed, chartaceous-membranous, 7-nerved, 
awned from one-fourth above the base; callus densely bearded with creamy hairs up to 3 mm 
long; awn 7.5-8.5 mm long, slightly bent, scaberulous; palea oblong, nearly as long as the 
lemma, 2-lobed, with the lobes acute, 2-keeled, with the keels ciliate, membranous; anthers 
up to 1.7 mm long. Upper lemma ovate-oblong when opened out, 4.5—5 mm long, 2-lobed, 
with the lobes unequally and acutely 2-toothed, membranous, 5-nerved, awned from the 
lowest third; callus densely bearded with hairs 2-3 mm long; awn up to 5 mm long, straight, 
scaberulous; palea narrowly oblong, nearly as long as the lemma, 2-keeled, with the keels 
ciliolate; anthers 1.5 mm long. 

Endemic. First discovered on Gough Island by Wace in 1968. He found it growing at four 
separate localities, all in wet heath communities beside streams or in wet gullies. It is 


* Material collected under Christophersen 826 is much smaller than the other numbered gatherings but agrees in 


other respects. 


392 E. W. GROVES 


distinguished from the other Tristan da Cunha island group species by its taller, stouter, 2-5- 
nhoded culms. 
Gough: Wace G.217 (K), G.223 (K), G.245 (K -holotype, BM), G.246 (K). 


Deschampsia wacei C. E. Hubbard, sp. nov. (Fig. 30) 
a D. mejlandii C. E. Hubbard, ligules brevioribus, laminis plerumque brevioribus angustioribus 
junciformibus involutis supra glabris, paniculis angustioribus, ramis laevibus differt. 

A densely caespitose perennial, 640 cm high, with numerous leafy innovations. Culms 
erect, slender, 1—2-noded in the lowest fourth, simple, smooth. Basal leaf-sheaths persistent, 
coriaceous, glossy, 14cm long, glabrous and smooth like the upper; ligules oblong- or 
ovate-lanceolate, acute, 2-10 mm long, membranous; blades juncaceous, rigid, curbed, 
1-12 cm long, 0.6-1.3 mm in diameter, involute, closely and prominently ribbed above, 
glabrous and smooth. Panicles lanceolate to oblong, erect or inclined, contracted, 2.5-15 cm 
long, 1-1.5 cm wide; rhachis glabrous, smooth; branches unequal, up to 3 cm long, smooth- 
pedicels 14 mm long, smooth. Spikelets oblong or gaping, 9-10 mm long, 2-flowered; 
rhachilla produced as a fine plumose bristle. Glumes narrowly oblong-lanceolate in profile, 
oblong opened out, acute to slightly obtuse, equal or nearly so, as long as the spikelet, keeled, 
1-3-nerved, glabrous, smooth, membranous. Lower lemma oblong-ovate when opened out, 
5-6 mm long, 2-lobed, with the lobes very unequally and acutely 2-toothed and the teeth 
mucronate, membranous, finely 7-nerved, awned, smooth except for the minutely 
scaberulous apex; callus bearded with white hairs 1-2 mm long; awn arising on the back of 
the lemma in the lowest third, up to 10 mm long, straight or nearly so, minutely scaberulous; 
palea nearly as long as the lemma, lanceolate-oblong, 2-toothed, 2-keeled, scaberulous on 
the keels, thinly membranous; anthers 1.2-1.4 mm long. Upper lemma oblong-ovate when 
opened out, 4-5 mm long, 2-lobed, with the lobes unequally and acutely 2-toothed, mem- 
branous, 7-nerved, awned; callus bearded with white hairs up to 1 mm long; awn arising in 
the lowest third of the lemma, 8-9 mm long, straight or slightly bent, minutely scaberulous; 
palea oblong, 3 mm long, 2-keeled, with the keels minutely scaberulous, membranous; 
anthers 0.7 mm long. 

This grass has been named in honour of Dr Nigel M. Wace, botanist on the 1955 Gough 
Island Scientific Survey Expedition. 

Endemic. First found on Gough Island by Wace during his visit there in December 1955, 
thereafter in January 1956 and again in 1968. It grows at four separate localities, in bryo- 
phyte associations on wet peat or beside streams between 500-620 m. This species 1s 
characterised by its habit of forming small dense tufts of leafy innovations. 

Gough: Wace (BM -holotype), 84, (BM), 118 (BM), G.218 (K), G.245 (K), G.302 (K). 


Digitaria sanguinalis (L.) Scop. 
Introduced. First collected on Tristan da Cunha near the Settlement by Stableford during 
1953-54. It was found again in the same locality by Dickson in 1962, and by Wace in March 
1968. A second locality for the species was discovered at the Potato Patches on Settlement 
Plain by Wace in March 1968. It is almost certain to have been introduced with imported 
fodder. 

Tristan: Dickson s.n. (BM, K); Stableford 7a (K), 7b (K), 46 (BM); Wace T.128 (K), T.161 (K). 


Echinochloa crus-galli var. breviseta (Doell) Neilr. 
Introduced. Collected by Stableford during 1953-54. Possibly introduced with imported 
fodder, along with Digitaria sanguinalis. 

Tristan: Stableford 3 (K), 45 (BM). 


Eleusine indica subsp. africana (O’Byrne) Phillips 

Eleusine africana O’Byrne 
Introduced. Collected on Tristan da Cunha in February 1962, growing on bare ground at the 
Settlement. 

Tristan: Dickson 122 (BM, K). 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 393 


—=—_— 


P= 
== 


ESSE 
SSsQ 


ed 
——=s 


Fig. 30 Deschampsia wacei C. E. Hubbard. a. habit (x 3); b. ligule (x 4); c. spikelet (x 4); d. glumes 
(x 4); e. floret (x 5); f. lemma (x 4); g. palea (x 4); h. stamens and ovary (x 7). All drawn from 
Wace 21. 

Fig. 31 Glyceria insularis C. E. Hubbard. a. habit (x 4); b. leaf tip (x 4); c. ligule (x 2); d. glumes 
(x 7); e. floret side view (x 7); f. floret opened side view (x 3); g. lemma (x 7); h. pales (x 7); i. ovary 
and stamens (x 8). All drawn from Mejland 1365. 


394 E. W. GROVES 


Festuca arundinacea var. mediterranea Hackel ex Battend & Trabut 
F. elatior subsp. arundinacea var. geniuna subvar. mediteranea Hackel 
Introduced. Collected on Tristan da Cunha in December 1976, in the school enclosure at the 
Settlement. It was probably introduced with imported mixed American fescue pasture seed. 
Tristan: Wace T.333 (K). 


Festuca rubra subsp. commutata var. barbata (Hackel) Howarth 

Introduced. First collected on Tristan da Cunha by Mejland in December 1937, being found 

in grassland by the Settlement. It was found again at the Settlement in 1962 by Dickson. 
Tristan: Mejland 102a (O, K), 104 (O, K, BM); Dickson 123 (K). 


Glyceria insularis C. E. Hubbard, sp. nov. (Fig. 31) 

a G. fluitanti (L.) R.Br., foliorum laminis apice pungentibus cartilagineis, lemmatibus plerumque 
obtusis vel minute trilobis vel abrupte acutis; glumis longioribus, paleis lemmatibus multo brevioribus, 
antheris minoribus, differt. 

Perennial, 25-55 cm high, rhizomatous, forming small tufts; shoots more or less com- 
pressed, 3-6 mm thick towards the base. Culms erect or ascending, closely sheathed, slender, 
simple. Leaves glabrous, green, sheaths, overlapping, smooth, striate and with scattered 
cross-nerves, the ribs obscurely vesiculate, ligules membranous, up to 10 mm long, 
becoming lacerate; blades linear, abruptly pungent-acute, with the tip cartilaginous, 
6-21 cm long, 4-10 mm wide, folded, at length flat, smooth, ascending, firm with yellowish 
cartilaginous margins. Panicles linear to lanceolate, erect or slightly inclined, somewhat 
secund, dense, 0.5—1.5 cm wide, 12-24 cm long; rhachis slender, smooth, striate; branches 
erect, 2—3-nate, up to 5cm long, 1-4-spiculate, smooth; pedicels 2-4 mm long, smooth. 
Spikelets linear-lanceolate to linear-oblong, 15-23 mm long, 2-3 mm wide, 6-10 flowered, 
green. Glumes acute or obtuse, membranous, |-nerved, smooth; lower lanceolate to 
narrowly oblong-lanceolate, 3.5-7.5 mm long; upper oblong-lanceolate, 5.5-8 mm long, 
acute or obtuse. Lemmas* oblong-lanceolate to lanceolate-oblong, obtuse, apiculate, mostly 
minutely 3-lobed, rarely abruptly acute, 6-7.5 mm long, closely and minutely scaberulous, 
coriaceous except for the hyaline tip; 7-nerved. Paleas elliptic-oblong to oblanceolate- 
oblong or oblong, very shortly 2-toothed or entire, 4-5 mm long, with narrowly winged 
keels. Lodicules connate, truncate, very small. Anthers 1.3-1.8 mm long. Grain elliptic- 
oblong, dorsally compressed, 2 mm long, dark-brown, hard. 

Endemic. It was first collected for the group in 1938, both on Tristan da Cunha and 
Inaccessible Island by Christophersen and Mejland during the visit of the Norwegian 
Scientific Expedition. The grass occurred mainly in damp grassy fields between 600-1000 m. 
It was found again on Tristan in 1968 by Wace by the crater-lake at Upper Cove, Gulch Hill, 
though nowhere abundant. Visiting Gough in 1956 Wace discovered it beside streams at two 
localities, i.e. in the crater of Edinburgh Peak and in the gullies of Gorydale. It was still 
present at these same sites when he was there in May 1968. 

Tristan: Christophersen 1267 (O, K); Mejland 1365 (O, K -holotype), 1546 (O, K); Wace T.243 (K). 
Inaccessible: Christophersen 2347 (O, K), 2465 (O, K). Gough: Wace 135 (BM), 143 (BM), G.238 (K). 

The two collections by Christophersen from Inaccessible Island cited above were at first 
considered sufficiently distinct to warrant description as a subspecies of G. insularis; but as 
this island has a large population of sea-birds, the greater vigour of the plants growing there is 
no doubt due to the highly nitrogenous nature of the soil. These two gatherings have taller 
culms 70-100cm high, with thicker vegetative shoots 10-15 mm wide at the base; 
leaf-blades up to 40cm long and 9-15 mm wide; ligules up to 16 mm long; panicles 
24-36 cm long, 1.5—2 cm wide; branches many-spiculate, with the longer ones divided into 
short appressed branchlets. It is of interest to note here that on these two G/yeria numbers of 
Christophersen (2347 and 2465), the following fungi were identified by E. Miiller of Ziirich: 
Anthostemella tumulosa (Roberge) Sacc. (2465), Ascochyta phleina R. Sprauge (2347), 
Gaeumannomyces graminis (Sacc.) v. Arx & D. L. Olivier (2465), Hendersonia culmicola 
var. minor (Sacc.) Sacc. (2465), and H. culmiseta Sacc. (2465). 


* The only South American species of G/yceria. G. multiflora Steud., has shorter lemmas. 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 395 


Holcus lanatus L. 

Introduced. First collected on Tristan da Cunha by Bonomi in February 1904. Here it has 
now become very common throughout the island, and in 1955 Wace observed that it 
occurred in three main communities, viz. in grass heath above the Base, as a coloniser of bare 
soil and gravel, and in swampy ground, forming mats of vegetation over pools in gullies and 
deep turf in level, swampy areas. 

Tristan: Bonomi 6 (NBG [herb. Mus. Austro-Afric. 1123], K), 7 (K, NBG [herb. Mus. Austro-Afric. 
1123 bis]); Christophersen 5 (O), 289 (O, BM, K), 647 (O), 830 (O); Dickson 80 (BM, K); Dyer 3525 
(PRE, NBG, BM), 3570 (PRE, NBG); Keytel 1833 (NBG, K, BM); Mejland 139 (O, BM, K); Stableford 
14 (K), 35 (BM), 38 (BM), 85 (sterile) (BM); Wace T.28 (BM), T.255c (K). Inaccessible: Christophersen 
2594 (O, K); Wace 1.2 (K). Nightingale: Christophersen 2098 (O, BM); Wace N.20 (K). Gough: van der 
Merwe 69 (PRE, K); Wace 52 (BM). 124 (BM). 


Hordeum glaucum Steudel 

Introduced. Found in 1953-54. Most probably it was introduced with animal fodder 

imported from Cape Town, South Africa, where it has been naturalised for many years. 
Tristan: Stableford | (K), 43 (BM). 


Hordeum leporinum Link 
Introduced. Collected in 1968 beside the bull-pen in an old fenced garden west of Big Watron 
near the Settlement on Tristan da Cunha. 

Tristan: Wace T.262 (K). 


Lolium x hybridum Hausskn. SHORT-ROTATION RYE-GRASS 
L. perenne x multiflorum 
Introduced. First found by Stableford in 1953-54. Obviously an introduction in imported 
seed mixture. 
Tristan: Stableford 9 (K); Wace T.158a (K). 


Lolium multiflorum Lam. 
Introduced. Found in 1953-54. Obviously an introduction with imported seed mixture. 
Tristan: Stableford 42 (BM), s.n. (K). 


Lolium perenne L. 
Introduced. First collected by Mejland in 1937 on Tristan da Cunha, where it was growing in 
moist pasture at the Potato Patches. It was found again in 1962 by Dickson at the side of a 
track near the west end of the Settlement. 

Tristan: Dickson 87 (BM); Mejland 160 (O, BM, K); Stableford s.n. (K). 


Lolium rigidum Gaudin 
Introduced. Found on Tristan da Cunha in 1968 by Wace who discovered it associated with 
other aliens (notably Trifolium subterraneum) in a levelled and seeded playing field north- 
east of the Settlement. 

Tristan: Wace T.158 (K). 


Parodiochloa C. E. Hubbard, gen. nov. 
a Poae L. foliorum laminis tenuiter acutis, lemmatibus cuspidatis vel aristatis, stigmatibus elongatis 
breviter pulverulentibus differt. 

Spikelets very numerous, oblong, becoming obcuneate, laterally compressed, cuspidate or 
awned, short-pedicelled on the branches of very dense panicles; rhachilla minutely and 
obscurely hairy or glabrous, disarticulating horizontally beneath each floret. Florets 2-4, 
¢; rhachilla produced and bearing a vestigial floret; callus extremely short, truncate, 
glabrous. Glumes persistent, subequal or the lower slightly shorter, keeled, acute, herbaceous 
about the keel, otherwise hyaline-membranous, |-3-nerved lower linear to narrowly oblong- 
lanceolate: upper narrowly lanceolate to narrowly oblong-lanceolate. Lemmas exceeding the 
glumes, contiguous, linear-lanceolate in side view, narrowly oblong or narrowly oblong- 
lanceolate when opened out, abruptly acutely acuminate, obtuse, truncate or slightly 
2-lobed, 5-nerved, with the lateral nerves evanescent in the upper third and with the middle 


396 E. W. GROVES 


nerve passing into a prominent cusp or straight relatively stout scaberulous awn up to two- 
thirds the length of the lemma, thinly herbaceous except for the narrow membranous- 
hyaline margins and apex, minutely hispidulous or scaberulous. Paleas up to four-fifths the 
length of the lemmas, narrowly oblong, 2-keeled, concave between the minutely ciliolate or 
scaberulous keels, hyaline-membranous. Lodicules 2, oblong or obcuneate, 2-toothed or 
lobed, hyaline, glabrous. Stamens 3; anthers narrowly oblong. Ovary glabrous; styles 2, 
terminal, close together; stigmas elongated, very slender, minutely and loosely hairy, 
exserted from the apex of the floret. Caryopsis lanceolate, enclosed but free between the 
slightly hardened lemma and palea, keeled on the back, shallowly grooved or flattened 
adaxially, with a narrow membranous apical appendage; embryo about one-fourth the 
length of the caryopsis; hilum punctiform, basal. Perennial, forming large dense tussocks or 
clumps; culms erect, striate; leaf-sheaths compressed, keeled, closed (but splitting); ligules 
membranous, glabrous; blades elongated, folded in bud, at length linear and flat, long- 
attenuated to a fine point; panicles spike-like, continuous or lobed or interrupted. 
Monotypic. 

Type species: Parodiochloa flabellata (Lam.) C. E. Hubbard. 

The genus Parodiochloa may be distinguished from Poa by the finely pointed tips of the 
leaf-blades, the cuspidate or awned lemmas, and the elongated puberulent stigmas, which are 
exserted from the tip of the floret. Typical species of Poa possess hooded tips to the leaf- 
blades, obtuse to acute or acuminate, very rarely mucronate, tips to the lemmas, the latter 
usually being hairy on the nerves below the middle and often with a flake of woolly hairs at 
the base, and plumose bushy stigmas which are laterally exserted towards the base of the 
floret. Poa cookii is somewhat similar to Parodiochloa in its densely spike-like panicle, but, 
except for its mucronate-tipped lemmas, it has all the characteristics of a species of Poa. 


Parodiochloa flabellata (Lam.) C. E. Hubbard, comb. nov. 

Festuca flabellata Lam., Encycl. Méth. Bot. 2: 462 (1788).—Poa flabellata (Lam.) Rasp. in Ann. Sc. 

Obsery. 2: 76, 78 (1829); J. D. Hook. in Phil. Trans. R. Soc. 168: 22 (1879); Skottsb. in K. svenska 

Vetensk Akad. Handl. 50(3): 14 (1913); Dallimore in Kew Bull. 1919: 217 (1919); W. Davies, 

Grasslands Falkland Is.: 48 (1939).—Dactylis caespitosa G. Forster in Commentat. Soc. Scient. 

gotting. 9: 22 (1789); Hook in Lond. J. Bot. 2: 298, tab. 10(1843); J. D. Hook., Fl. Antarct. 2: 384, 

tab. 136, 137 (1847). 

Occurs in temperate South America, Falkland Islands, Fuegia (especially on treeless 
islands), South Georgia and Gough Island, and has also been introduced into the Shetland 
Islands (United Kingdom). It grows in the neighbourhood of the sea, on coastal cliffs and on 
mountain slopes. It is said to possess valuable feeding properties and is greedily eaten by 
grazing animals. 

Introduced (?), As no collections of this grass seem to have been made on the group prior to 
this century, it is now thought unlikely to have ever occurred as native on Tristan da Cunha, 
Inaccessible and Nightingale Islands (Wace & Holdgate, 1976: 45). Had it been once as 
prominent on them as it is on Gough Island today, it seems unlikely that it would have been 
missed by the early visiting naturalists, particularly as the northern islands of the group were 
the more frequented by sailing vessels. 

P. flabellata is now the dominant tussock grass on Gough Island, forming dense and pure 
communities on the western and northern cliff slopes, and also above the cliff tops on some 
of the more level ground below 300 m (Figs 20 and 32). It is also the principal dominant in 
some sheltered hollows at about 490 m to the south and west of Gonydale (Wace, 1961: 
341-342). It is confined to areas of good drainage and the bottoms of hollows, although its 
distribution is in places clearly related to the extent of the penguin rookeries (Wace & 
Dickson, 1965: 307). 

Gough: van der Merwe 30 (PRE, K); Wace 46 (BM), 87 (BM), G.207 (K). 


Paspalum dilatatum Poiret 
Introduced. First found on Tristan da Cunha by Dickson in February 1962 at the Settlement. 
It was collected there again by Wace in 1968, to the west of Big Watron amongst the New 


397] 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 


“Q0BM [SIN :ydeis0}0Yyd 896] Av ‘Avg 1seNd 
JAOQP S]JI[9 9Y1 JO SAdOJs JY] SIQAOD ‘DUDIJADYIIMADI “eA DS1INUI Sida] do1siyy Aq payeurwop ‘ysnq 
UlaJ ‘VALOGAD DIIAY_ JO SII] Maj & AQ P3JUNOULINs se puR YOY ABMYIIY IYIO[D DaIvUIpUNAD 
puljipdgy puke vjIpjjaqv]f vo]YyIoIppADg Sosseis y) JO sdwinj[D “Jeol ay} 0} JUIOg punoljney 
pue ‘rd eo pee ee oY} Ul YOOY AeMYOIY Bue ped udg][H YIM Pues sear ZE Sly 


398 E. W. GROVES 


Zealand flax (Phormium tenax), where it formed a thick growth in a sheltered situation, 
protected from grazing. 
Tristan: Dickson 66 (K); Wace T.109 (K), T.180 (K), T.338 (K). 


Pennisetum clandestinum Hochst. ex. Chiov. 
Introduced. Collected on Tristan da Cunha in November 1976, where a clump was found 
colonising sand near the lagoon east of the new lava. Wace also noticed the species a month 
later forming conspicuous patches in newly sown pastures near the Settlement. 

Tristan: Wace T.301 (K), T.321 (K). 


Phalaris tuberosa L. 
Introduced. Collected on Tristan da Cunha in February 1962 by Jenny’s Watron on Settle- 
ment Plain. 

Tristan: Dickson 81a (K). 


Poa annua L. 

Introduced. First collected on Tristan da Cunha by Macgillivray in 1852. Here it is frequent 
beside the lake in the Upper Crater by Cave Gulch, and has now also become a common 
weed at the Potato Patches and areas of disturbed ground below 1220 m. On Inaccessible 
Island it has been found about Blenden Hall, and on Nightingale Island along a trail path. On 
Gough Island it is frequent in wet peaty situations particularly near nests of the gony or 
Tristan subspecies of the wandering albatross (Diomedea exulans dabbenena) and around 
wallows of the elephant seal (Mirounga leonina) near to huts. 

Tristan: Christophersen 9 (O, BM), 12 (O), 401 (O), 444 (O, K), 541 (O), 550K (O); Dyer 3530 (PRE, 
NBG, K); Macgillivray 344 (K); Mejland 133 (O); 1156 (O, BM, K); Moseley s.n. (BM, K); Siggeson 53 
(O); Stableford 7a (K), 50 (BM); Wace T.50 (BM), T.245 (K), T.255b (K). Inaccessible: Christophersen 
2384 (O), 2482 (O, BM, K); Moseley s.n. (BM, K). Nightingale: Christophersen 2157 (O); Wace N.61 
(K). Gough: Brown s.n. (K); Christensen s.n. (O); van der Merwe 73 (K, PRE); Wace 44 (BM, G.201 
(K), G.215 (K). 


Poa infirma Kunth 

Poa exilis (Tomm.) Murb. 
Introduced. The presence of this species is based on a single gathering made on Tristan da 
Cunha by Mejland in January 1938 which was subsequently determined by Axel Nannfeldt. 
It was found growing at about 1000 m in the Hottentot Gulch. This widespread annual grass 
is a native from the Canary Islands, through the Mediterranean region and north Africa to 
north-west India. It has been introduced in other countries, including the British Isles, the 
Channel Isles, and South America. 

Tristan: Mejland 1361 (O). 


Poa pratensis L. 

P. pratensis L. subsp. pratensis 
Introduced. First collected on Tristan da Cunha by Moseley in 1873. It is a common grass 
there in more or less open communities at lower levels all over the island, especially in 
grassland near the Settlement and at the Potato Patches. Elsewhere it is often found 
colonising screes and loose soil in gullies. On Gough Island it is common around the huts 
and by the lower reaches of Glen Stream. 

Tristan: Christophersen 63 (O, BM), 206 (O, BM, K), 294 (O), 539 (O, BM); Dickson 162 (BM); 
Keytel 1837 (NBG, K); Mejland 167a (O), 180 (O), 1356 (O, BM, K); Moseley s.n. (BM, K, E); Wace 
T.27 (BM), T.51 (BM). Gough: Wace 8 (BM). 


Poa subcaerulea Sm. 
P. pratensis subsp. subcaerulea (Sm.) Tutin 


Introduced. This taxon is sometimes included as either a subspecies or variety of Poa 
pratensis, but Hubbard (1954: 169) considers it sufficiently distinct to be regarded as a 
separate species. The earliest collection was made on Tritan da Cunha in December 1937 by 
Mejland, who found it in pasture near the Settlement. 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 399 


Tristan: Dickson 21 (BM, K), 88 (BM); Dyer 3625 (PRE, K), s.n. (K); Mejland 103 (O, K); Stableford 
11 (K), 84 (BM), 89 (BM), 116 (BM); Wace T.349 (K). 


Poa trivialis L. 
Introduced. First found by Mejland on Tristan da Cunha in 1937 (determined by Nannfeldt 
and cited by Hafsten, 1951: 14), growing in pasture at the Settlement. It was found again at 
the Settlement in December 1976 by Wace, growing amongst New Zealand flax. 

Tristan: Mejland 124 (O); Wace T.340 (K). 


Polypogon mollis (Thouars) C. E. Hubbard & E. W. Groves, comb. nov., 

Pharlaris mollis Thouars, Equisse Flore Il'Isle de Tristan d’Acugna: 37 (1808).—Polypogon 

intermedius Carmich. in Trans. Linn. Soc. Lond. 12: 504 (1818). 
A rhizomatous perennial, up to 75 cm high, rhizomes slender, spreading; culms relatively 
stout, clustered, geniculately ascending, simple, several-noded. Leaf-blades linear, up to 
25 cm long and 8 mm wide, smooth or nearly so; ligules oblong, becoming lacerate, up to 
6mm long. Panicles continuous, or interrupted and + lobed, very densely spiculate, 
10-15 cm long, 1.5-2 cm wide; pedicels articulated at the base, 2-6 mm long, scabridly 
hispid. Spikelets continuous with and falling with the pedicels at maturity. Glumes stiffly 
and minutely hispid, equal or unequal, awned from the tips 5-7 mm long (including the 
awn), lanceolate, acutely acuminate, membranous, |-nerved, the awn fine, scaberulous, up 
to 2.5mm long, the upper glume shorter and slightly narrower. Lemma ovate, obtuse, 
2-2.5 mm long, finely 5-nerved, with the lateral nerves minutely excurrent, aristulate and 
scabrid, the rest glabrous and smooth, awned from just below the tip, with the awn straight, 
0.6-1.5 mm long; palea oblong, obtuse, | mm long, hyaline, nerveless; anthers 0.3-1 mm 
long; grain | mm long, elliptic-oblong, dorsally compressed. 

Endemic. First collected for the group on Tristan da Cunha in 1793 by Aubert Du Petit- 
Thouars who published it (loc cit.) in his list under the name Phalaris mollis. A re- 
examination of this type material in Paris together with the original description shows that 
this grass belongs to the genus Polypogon. Carmichael, visiting Tristan da Cunha in 
1816-17, found what he considered to be a new species of Polypogon which he published as 
P. intermedius. However, the type specimen is identical with that of Thouars’ Phalaris 
mollis. The precise habitat details and localities are unknown for the only two specimens so 
far collected on Tristan, but the plants gathered on Inaccessible Island by Christophersen 
(February 1938) were found growing in a swampy depression near the western end of the 
plateau. 

Tristan: Carmichael s.n. (BM -holotype of P. intermedius); Thouars s.n. (P -holotype of P. mollis). 
Inaccessible: Christophersen 2466 (O, K). 


Polypogon monspeliensis (L.) Desf. 
Introduced. Found on Tristan da Cunha by Keytel during 1908-09, and not found there 
again until 1968, when it was collected by Wace at the margins of freshwater swamps to the 
east and west of Pitbite, and also near the eastern edge of the new lava, growing amongst 
drifting beach sand and pebbles. It is a widespread grass in Europe, Africa and Asia and was 
introduced to Tristan da Cunha years ago, probably from South Africa. 

Tristan: Keytel 1835 (NBG, K); Wace T.154 (K). 


Spartina arundinacea (Thouars) Carmich. TUSSAC or TUSSOCK GRASS 
Ponceletia arundinacea Thouars 
Native. First collected on Tristan da Cunha by Aubert Du Petit-Thouars in 1793. This grass 
forms a tussock community on all of the islands, but on Tristan da Cunha its former native 
distribution along some of the coastal strips has been much reduced through clearing and 
firing and through the grazing of animals. Tussac was also cultivated along with flax on the 
inside of the walled cottage gardens, mainly to protect growing plants and vegetables from 
the strong winds (Christophersen, 1940: 26). From sea-level to about 100 m on that island it 
has been replaced by secondary grassland consisting of introduced species. In 1968 Wace 
found that on Inaccessible Island it was dominant both on the steep cliffs and on the more 
gently sloping ground burrowed by birds. On Nightingale Island in the same year he found 


400 E. W. GROVES 


that it was dominant over most of the island forming ‘a bamboo-like dense cover with 
individual plants 2—2.5 m in height’. Earlier, in 1955, on Gough he found the tussocks to a 
height of 3 m and with a circumference of 1.5 m, growing along the shores and in clearings in 
the Phylica forest. 

Tristan: Christophersen 288 (O, BM, K), 425 (O, K); Carmichael s.n. (BM); Dickson | (K), 30 (K), 99 
(BM); Dyer 3576 (PRE, K), 3577 (PRE, NBG, K); Macgillivray 345 (K), s.n. (BM); Mejland 140 (O), 
1170 (O), 1479 (O); Milne s.n. (K); Moseley s.n. (K, BM); Siggeson 56 (O); Stableford s.n. (K); Swain in 
Rogers s.n. (K); Thouars s.n. (P -holotype); Wace T.22 (BM). Inaccessible: Christophersen 2595 (O, 
BM, K); Moseley s.n. (K, E); Wace 1.13 (K), 1.14 (K). Nightingale: Christophersen 2167 (O, BM); 
Stableford 4 (K), 4a (K), 73 (BM); Wace N.13 (K). Middle: Christophersen 2020 (O, BM, K). Gough: 
Brown s.n. (K, E); Christensen s.n. (O); Fleming 36 (E); van der Merwe 60 (PRE, K); Wace 43 (BM). 


Sporobolus africanus (Poiret) Robyns & Tournay HARD TWIT 
S. capensis (Willd.) Kunth; S. indicus auct., non (L.) R.Br. 


Introduced. First collected on Tristan da Cunha by Keytel, 1908-09. Since that date it has 
become abundant in pastures at the Settlement and on the lower slopes up to Burntwood. 

Tristan: Christophersen 270 (O), 1322 (O); Dickson 3 (BM), 82 (K); Dyer 3541 (NBG, K), 354la 
(NBG, K); Keyte/ 1836 (K); Mejland 1301 (O, BM, K), 1304 (O), 1305 (O), 1672 (O, BM, K); Stableford 
5 (K), 47 (BM), 90 (BM); Wace T.88 (BM), T.110 (K), T.339 (K). 


Vulpia bromoides (L.) Gray 

Introduced. First collected on Tristan da Cunha by Macgillivray in November 1852. 
Christophersen (1937: 8) says of this widely distributed grass that ‘it is probably of late intro- 
duction to Tristan da Cunha’. On Macgillivray’s visit it was most likely to have been found at 
the Settlement, where, in the pastures, it has since been collected several times. It has also 
occurred at the Potato Patches, Seal Bay, Stony Beach and Hottentot Gulch. More recently, 
in 1968, Wace noticed it colonising the new volcano and growing on waste ground near the 
fish factory. It has also been found on Inaccessible Island where it was growing on talus above 
Blenden Hall. 

Tristan: Christophersen 222 (O), 1132 (O), 1792 (O); Dickson 17 (K), 92 (BM), 166 (K); Dyer 3533 
(NBG, K), 3533 bis (PRE); Keytel 1839 (NBG, K, BM); Macgillivray s.n. (K); Mejland 102 (O, BM), 
109a (O); 113 (O, K), 1362 (O), 1418 (O), 1427 (O); Milne s.n. (K); Moseley s.n. (K); Siggeson 54 (O*); 
Stableford | (K), 1a (K), 1b (K), 33 (BM), 83 (BM), 87 (BM), 108 (BM); Wace T.120 pro parte (K), 
T.124 (K), T.126 (K), T.313 (K). Inaccessible: Christophersen 2485 (O). 

[Vulpia myuros (L.) C. C. Gmelin 

Festuca myurus L. 

This species is listed by Christophersen (1937: 16) as being from Tristan da Cunha. He does not 
elaborate, but it is no doubt based on Festuca myurus given in the list of introduced plants observed by 
Moseley during the visit of HMS Challenger in October 1873 (Hemsley 1885: 145). C. E. Hubbard 
(pers. comm.) considers Moseley’s voucher specimen, which is at K, to be Vulpia bromoides (q.v.).] 


CONIFERAE 


Pinus caribaea Morelet 

Introduced. Wace & Dickson (1965: 333) mention the presence on Inaccessible Island of 
three trees situated behind the ruined cottage at Saltbeach. The trees appeared not to be 
regenerating when examined in 1968 (Wace & Holdgate 1976: 44-7). Wace & Holdgate 
(1976: 45) mention the existence of a pine plantation on Tristan da Cunha near Sandy Point 
but do not indicate the identity of the species. f 


PTERIDOPHYTA 
The arrangement for the families and genera follows Crabbe, Jermy, & Mickel (1975). 


* Not seen in Oslo herbarium in October 1973, but recorded as being there by Christophersen (1937: 8). 

+ Anexperimental introduction of some alien trees and shrubs to Tristan da Cunha, made in 1937, including Pinus 
canariensis, P. halepensis, P. insignis, and P. pinaster did not prove successful. Christophersen collected an 
herbarium specimen (No. 1785, now in O) of P. radiata on Tristan da Cunha in March 1938 from a tree in Gordon 
Glass’s garden at the Settlement, obviously an introduction; it is not known if this tree survives. 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 401 
LYCOPODIACEAE 


Lycopodium diaphanum (Beauv.) Swartz DEVIL’S FINGERS 

Lepidotis diaphana Beauv.; Lycopodium clavatum sensu Thouars, non L. 

Endemic. First collected on Tristan da Cunha by Aubert Du Petit-Thouars in 1793. On 
Tristan it occurs mainly amongst fern scrub and in the Phylica forest up to 650 m, although 
it is locally common in the Empetrum/Blechnum heath at lower altitudes. 

Tristan: Carmichael s.n. (BM); Christophersen 207 (QO), 455 (O, BM, K), 529 (O), 1032 (O); Dickson 
24 (BM), 151 (BM, AAS); Dyer 3563 (PRE, NBG); Keytel 1840 (NBG, BM); Glass s.n. (BM); Keytel s.n. 
[herb. Marloth 4701] (PRE); Herb. Menzies s.n. (BM); Mejland 1820 (O); Richardson s.n. (O); Rogers 
1 (BM); Stableford 58 (BM); Thouars s.n. (P* -holotype of Lepidotis diaphana),; Wace T.4 (BM), T.5 
(BM). Inaccessible: Christophersen 2306 (O, BM, K); Stableford 123 (BM). Gough: van der Merwe 27 
(PRE); Wace 101 (BM), 158 (BM). 


Lycopodium magellanicum Swartz 

Native. First collected on Tristan da Cunha by Christophersen in 1937, where it is common 

in wet peat in bogs and on mossy slopes in the Phylica forest. It has also been found on 

Gough Island, where it is frequent on blanket peat at about 300 m throughout the island. 
Tristan: Christophersen 454 (O, BM, K), 1019 (O); Dickson 175 (BM); Mejland 1167 (O, BM); Wace 

T.1 (BM). Gough: Wace 19 (BM), 147 (BM). 


Huperzia insularis (Carmich.) Rothm. 

Lycopodium insulare Carmich.; Urostachys insularis (Carmich.) Herter; Lycopodium saururus 
Hemsley 
Endemic. First collected on Tristan da Cunha by Carmichael in 1816. It is common on the 
islands in wet peat in bogs up to 760 m, sometimes forming clumps. A specimen collected by 
Christophersen on Inaccessible Island (No. 2356) and assigned by Christensen (1940: 23) to 
L. selago var. hessei Herter may possibly belong here. 

Tristan: Carmichael s.n. (K -holotype of L. insulare); Christophersen 1021 (O, BM, K), 1066 (O), 
1273 (O, BM, K); Mejland 1176 (O, BM); Wace T.2 (BM), T.3 (BM). Inaccessible: Christophersen 
2356 (O, BM, K), 2475 (O); Stableford 122 (BM). Nightingale: Christophersen 2214 (O); Crosbie s.n. 
(E); Moseley s.n. (K, -holotype of L. saururus, E). Gough: van der Merwe 36 (PRE); Wace 20 (BM), 102 
(BM), 159 (BM). 


OPHIOGLOSSACEAE 


Ophioglossum opacum Carmich. 
Endemic. First collected on Tristan da Cunha by Carmichael, 1816-17, ‘high on the dome’ 
[presumably the Peak]. It has not been collected on Tristan since, but in 1956 it was found by 
Wace in three different localities on Gough Island, in wet peat amongst byrophytes, all above 
600 m. 

Tristan: Carmichael s.n. (BM -holotype). Gough: Wace 105 (BM), 146 (BM), 150 (BM). 


ADIANTACEAE 
ADIANTOIDEAE 


Eriosorus cheilanthoides (Swartz) A. F. Tyron 

Grammitis cheilanthoides Swartz; Gymnogramma cheilanthoides (Swartz) Kaulf.; Asplenium 
filipendulaefolium Thouars; G. filipendulaefolia (Thouars) Hook. 
Native. First recorded on Tristan da Cunha by Aubert Du Petit-Thouars in January 1793, 
who published it as new under the name Asplenium filipendulaefolium in the text, but as 
Grammitis cheilanthoides in the caption to the plate (Aubert Du Petit-Thouars, 1808; 34 & 
tab. IV). It has since been found on that island on several occasions, and also on Inaccessible 
and Gough Islands, occurring on peaty rock ledges between 90-400 m. it exhibits consider- 
able variation in size and leaf form, but clearly has a South American affinity (cf. Tyron, 
1970: 129-133). Spores of this species were recovered by Hafsten from peat cores collected 


* Not seen on my visit in September 1975. 


402 E. W. GROVES 


by N. M. Wace on Tristan da Cunha and Gough Island in 1955-56 (Hafsten, 1961). Tyron 
(1970: 61) commenting on his work considers that this fern has probably existed on the 
Tristan group for at least 5,000 years. The latter author is also of the opinion that Tristan 
examples of this fern ‘are not distinguishable from some occurring in the Andes formerly 
called E. elongatus (Grev. & Hook.) Copel. or E. flabellatus (Grev. & Hook.) Copel.’ (Tyron, 
1966: 271). 

Tristan: Carmichael s.n. (BM, K, E [ex. herb. Menzies], E [ex. Wernerian Soc. volume]); 
Christophersen 202 (O), 210 (O, BM, K), 456 (O, BM); Dickson 158 (BM*); Dyer 3578 (PRE); Keytel 
1847 (NBG); Thouars s.n. (P -holotype of A. filipendulaefolium, BM); Wace T.36 (BM). Inaccessible: 


Christophersen 2378 (O), 2502 (O), 2554 (O); Stableford 129 (BM). Gough: van der Merwe 17 (PRE), 
21 (PRE); Wace 75 (BM), 104 (BM), 137 (BM). 


Adiantum poiretii Wikstr. 

A. aethiopicum sensu Thouars, non. L.; A. crenatum Poiret (1826), non Willd. (1826) 

Endemic. First found on Tristan da Cunha by Aubert Du Petit-Thouars in 1793 and 
recorded under the name A. aethiopicum. It occurs on the islands in gullies and crevices or 
under rock ledges in damp grass, sometimes where there is dripping water. 

Tristan: Bonomi s.n. (NBG [herb. Mus. Austro-Afric. 1113]); Carmichael s.n. (BM, K); 
Christophersen 212 (O), 429 (O, K), 537 (O, BM); Dickson 20 (BM); Dyer 3560 (K); Keytel 1827 (NBG, 
K, BM); Mejland 181 (O); Moseley s.n. (K, BM); Thouars s.n. (P-in herb Jussieu -holotype of A. 
crenatum and A. poiretii, BM); Wace T.94 (BM). Inaccessible: Christophersen 2386 (O); Moseley s.n. 


(K). Nightingale: Stableford 7 (K). Gough: Brown s.n. (E),; van der Merwe 53 (PRE), 75 (PRE); Wace 2 
(BM). 


VITTARIOIDEAE 


Vittaria vittarioides (Thouars) C.Chr. 

Pteris vittarioides Thouars; Vittaria stricta Carmich. 

Endemic. First collected on Tristan da Cunha by Aubert Du Petit-Thouars in January 1793 
and published as a new species. On Tristan it occurs mainly in the Blechnum palmiforme/ 
Empetrum/Phylica heath and in the Phylica forest, in which latter community it is also 
found on Gough Island. On Nightingale and Inaccessible Islands it has been found growing 
in gullies, in rock crevices and on rock ledges. 

Tristan: Carmichael s.n. (BM -holotype of V. stricta, E [ex herb. Menzies]); Christophersen 527 (O, 
BM, K); Dickson 169 (BM); Dyer 3559 (PRE); Mejland 618 (O), 1625 (O, BM, K); Thouars s.n. 
(P -holotype of P. vittarioides [ex herb. Desvaux and ex herb. Bory de St. Vincent]); Wace T.33 (BM). 
Inaccessible: Christophersen 2377 (O); Moseley s.n. (K, P). Nightingale: Christophersen 2001 (O, BM, 
K), 2179 (O), 2226 (O, BM). Gough: Brown s.n. (K, E); Wace 54 (BM). 


HYMENOPHYLLACEAE 


Hymenophyllum aeruginosum (Poiret) Carmich. 

Trichomanes aeruginosum Poiret 
Endemic. First collected on Tristan da Cunha by Aubert Du Petit-Thouars in January 1793. 
This small fern, the largest specimens having fronds between 10-15 cm long, occurs 
frequently amongst moss, at the bases of other ferns and on trunks of tree ferns, often in dense 
shade. Occasionally it has been noted as growing on the faces of damp cliffs. 

Tristan: Carmichael s.n. (K, BM, E); Christophersen 203 (O, BM), 245 (O, K), 452 (O, BM, K), 1245 
(O), 1270 (O); Dickson 100 (BM); Dyer 3358 (PRE, NBG, K), s.n. (PRE); Keyte/ 1825 (NBG, K); 
* There is also an herbarium sheet in BM prepared from material cultivated at Leeds University of Dickson’s 
original 1966 Tristan gathering. 


+ The name A. poiretii has, in the past, been applied to all forms of this taxon whether on the Tristan da Cunha 
islands or on continental Africa, with A. thalictroides as a synonym. Pichi-Sermolli (1957: 695), however, in a 
critical examination of numerous specimens, is of the opinion that the epithet poiretii should be reserved solely for 
the Tristan group taxon (where it is endemic) and that the continental African taxon (which has separate constant 
characters) should bear the epithet thalictroides. 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 403 


Macgillivray 336 (K), s.n. (BM); Milne s.n. (K); Moseley s.n. (K, BM); Thouars s.n. (P -holotype*); 
Wace T.12 (BM). Inaccessible: Christophersen 2345 (O), 2614 (O). Nightingale: Christophersen 2034 
(O, BM, K), 2211 (O). Gough: van der Merwe 74 (PRE); Wace 41 (BM), 64 (BM), 138 (BM). 


Hymenophyllum peltatum (Poiret) Desv.t 

Native. First collected on Tristan da Cunha by Siggeson in 1934 although possibly 
overlooked by earlier visitors. It is a frequent epiphyte on Phylica arborea, and is often found 
in association with the previous species on Blechnum palmiforme trunks. Fallen dead wood 
and peat banks in the Phylica scrub up to 500 m are also frequent habitats. 

Tristan: Christophersen 58 (O), 78 (BM), 204 (O), 246 (O, BM), 678 (O), 827 (O, BM, K), 1057 (O, 
BM), 1225 (O); Dickson 102 (BM), 157 (BM); Dyer 3582 (PRE, NBG); Mejland 301 (O), 817 (O); 
Siggeson 36 (O). Inaccessible: Christophersen 2479 (O), 2496 (O). Gough: van der Merwe 20 (PRE); 
Wace 42 (BM), 65 (BM), 139 (BM). 


Hymenophyllum tunbrigense (L.) Sm. 
Native. Based on cytological work carried out on living ferns brought back by J. H. Dickson 
from Tristan da Cunha in 1966, Manton & Vida (1968) discovered amongst them H. 
tunbrigense (2n=26) on the island group, as distinct from the long established populations of 
H. peltatum (2n=36). Dickson’s 1966 material had been found in Little Sandy Gulch on 
Tristan da Cunha, attached to the rhizomes of Elaphoglossum laurifolium growing on 
shaded vertical rocks at about 300 m. See also second footnote below. 

Tristan: Christophersen 460 (O, BM, K), 1244 (O, BM); Dyer 3582 pro parte (PRE, NBG, BM); 
Wace T.11 (BM). 


Trichomanes angustatum Carmich. 

T. tenerum sensu Hemsley, non Sprengel 
Native. First collected on Tristan da Cunha by Carmichael, 1816-17, who described his 
plant as a new species (Carmichael 1819: 513). On Tristan it occurs in Blechnum palmiforme 
scrub on rock and in shaded, damp places in gullies and ravines, and on Inaccessible Island 
on shaded, wet rocks, under Spartina arundinacea. Christensen (1940: 3) considers that this 
species is ‘without any near African relative but exceedingly close to 7. tenerum Spreng., a 
common species of tropical America (Mexico to south Brazil)’. 

Tristan: Carmichael s.n. (K -holotype of 7. angustatum); Dickson 32 (BM, AAS), 159 (BM, AAS); 
Keytel 1826 (NBG), s.n. (K); Macgillivray 335 (K); Mejland 191 (O, BM), 1149 (O, BM, K); Milne s.n. 
(K); Wace T.10 (BM). Inaccessible: Christophersen 2583 (O, BM, K); Dickson 142 (BM). Nightingale: 
Stableford 15 (K). 


GRAMMITIDACEAE 


Grammitis magellanica Desv. subsp. magellanica 

Polypodium magellanicum (Desv.) Sturm.; P. billardieri var magellanicum (Desv.) C.Chr., 
Grammitis billardieri var. magellanica (Desv.) de la Sota 
Native. First collected on Tristan by Carmichael in 1816, with further examples collected by 
Christophersen in 1937 and Wace in 1955. Wace also found it on Gough Island in 1956. It 
occurs in rock crevices and gullies and also on truncks of Blechum palmiforme and Phylica 
arborea. 

Tristan: Carmichael s.n. (K-2 sheets); Christophersen 543 (O, BM), 1058 (O, BM); Mejland 1172 (O); 
Wace T.65 (BM). Gough: Wace 96 (BM), 148 (BM). 


* Poiret in his original description in Lamarck’s Encyclopédia Méthodique 8: 76 (1808) states that ‘Cette plante a 
été recueillie par M. Borry de Saint-Vincent, dans I’Ile déserte de Tristan d’Acugna. (V.s. in herb du Petit-Thouars) : 
Actually the reverse is true in that this specimen was collected from Tristan da Cunha by Aubert Du Petit-Thouars 
and then passed into Bory de Vicent’s herbarium. from whence it later became incorporated in the collections at P. 

+ Christensen (1940: 56), in discussing the ferns collected during the Norwegian Scientific Expedition, 1937-38, on 
the Tristan da Cunha group recognised two varieties of H. peltatum. To the first of these, var. menziesii (Presl.) C. 
Chr. (based on H. menziesii Presl, known until then from southern Chile, Fuegia and Juan Fernandez), he assigned 
ten of the Christophersen and Mejland gatherings, and to the second variety, for which he proposed no name, he 
assigned Christophersen 460 and 1244 (both collected on Tristan da Cunha). Commenting upon the last two 
gatherings, Manton & Vida (1968: 366-7) considered that in the light of their cytological work on Tristan ferns ‘there 
seems no doubt that they should in fact have been referred to H. tunbrigense’. These two gatherings have been cited 
in this present paper under H. tunbrigense (q.v). 


404 E. W. GROVES 


Grammitis poeppigiana (Mett.) Pichi-Serm. 

Polypodium poepiggianum Mett.; G. australis var. nana Franch.; G. nana (Franch.) Brack.; G. 
billardieri var. magellanica forma nana (Franch.) de la Sota; Polypodium billardieri var. magellanicum 
forma nana (Franch.) Skottsb.; G. armstrongii Tindale 
Native. First recorded on Tristan by Christophersen in 1937 and later during the same 
expedition on Inaccessible Island and Nightingale Island. More recently (1956) it was 
collected on Gough by Wace, who noted that it was often depauperate. It occurs on the 
islands in similar habitats to G. magellanica. 

Tristan: Christophersen 201 (O), 522 (O, K), 543 (O, BM), 631 (O, BM); 825 (O, BM, K); Dickson 
120 (BM, AAS), 164 (BM); Mejland 172 (O), 1391 (O, BM). Inaccessible: Christophersen 2359 (O), 
2589 (O). Nightingale: Christophersen 2032 (O -sterile), 2180 (O). Gough: Wace 112 (BM), 141 (BM). 


Parris (in litt.) is of the opinion that Christophersen 201 ‘is an extremely large and most 
uncharacteristic poeppigiana: although the frond size is that of magellanica, as is the 
altitude, the spore size and habitat is of poeppigiana. Perhaps the specimen was collected 
from a very shaded and moist rock crevice.’ Of Wace 141 she considers it to be ‘an atypical 
poeppigiana because the habitat (above the forest line) is not of magellanica, which is almost 
always a forest epiphyte.’ 


DENNSTAEDTIACEAE 
DENNSTAEDTIOIDEAE 


Hypolepis rugosula var. villoso-viscida (Thouars) C.Chr. 

Polypodium villoso-viscidum Thouars; Cheilanthes viscosa Carmich. 

Endemic variety. First collected on Tristan da Cunha by Aubert Du Petit-Thouars in 1793 
and described as Polypodium villoso-viscidum. On Tristan da Cunha and Inaccessible Island 
it grows amongst the Blechnum palmiforme scrub, and on Gough Island Wace noted it as 
occurring in the Histiopteris association. Christensen (1940: 7) says that the variety ‘is very 
variable in size and shape...’. 

Tristan: Carmichael s.n. (BM, K -holotype of C. viscosa); Christophersen 526 (O, BM); Mejland 1158 
(O, BM, K); Rogers s.n. (K); Thouars s.n. (P -holotype of P. villoso-viscidum). Inaccessible: 
Christophersen 2498 (O, BM, K), 2503b (O), 2558 (O), 2621 (O). Nightingale: Christophersen 2178 (O), 
2224 (O, K), 2233 (O, BM, K); Dickson 135 (BM); Rogers s.n. (K). Gough: Wace 30 (BM), 128 (BM). 


Histiopteris incisa var. carmichaeliana (Agardh) C.Chr. 

Pteris vespertilionis var. carmichaeliana Agardh 
Endemic variety. First collected on Tristan da Cunha by Carmichael, 1816-17, the specimen 
being later decribed by Agardh (1839: 80) as a new variety in honour of the collector. On 
Tristan it occurs in the Blechnum palmiforme/Empetrum/Phylica heath, and on Gough 
Island it is common on the peat up to as far as the Phylica forest zone (Figs. 12 & 33). 

Tristan: Carmichael s.n. (K -holotype BM); Christophersen 461 (O, BM, K), 528 (O, BM, K), 1063 
(O); Dickson 176 (BM); Fleming 85 (E); Keytel 1840 (K), 1842 (NBG); Macgillivray 330 (K), s.n. (BM); 
Milne s.n. (K); Rogers s.n. (K); Wace T.41 (BM). Inaccessible: Christophersen 2555 (O); Moseley s.n. 
(BM, K, E). Nightingale: Christophersen 2004 (O, BM, K); Moseley s.n. (BM, K); Rogers s.n. (K); 
Stableford s.n. (K). G2ugh: MacMillan s.n. (PRE); van der Merwe 29 (PRE), 64 (PRE); Swain s.n. (L); 
Wace |1 (BM), 31 (BM). 


THELYPTERIDACEAE 


Amauropelta bergiana var. tristanensis Holttum 

Dryopteris tometosa auct., non (Thouars) Kuntze; Thelypteris tomentosa auct., non (Thouars) Ching 
Endemic variety. Aubert Du Petit-Thouars (1808: 32, 67), in his list of plants collected on 
Tristan on his way to the Mascarenes, published a new species Polypodium tomentosum. 
However, Holttum (1974: 134-5), after re-examining Aubert Du Petit-Thouars’ type in the 
Paris herbarium, is of the opinion that the specimen was possibly wrongly localised, as the 
true Polypodium tomentosum (now Amauropelta tomentosa (Thouars) Holttum) is other- 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 405 


wise known only from Mauritius and Réunion.* Holttum (1974: 133-4) considers the 
Tristan-Gough taxon to be allied to A. bergiana (Schlectend.) Holtt., of mainland Africa, and 
has, therefore made it a variety, var. tristanensis. Carmichael’s specimen, which would have 
been collected on Tristan da Cunha during 1816-17, is therefore the earliest authentic 
record. It occurs occasionally in the Blechnum palmiforme/Phylica scrub on Tristan da 
Cunha and Inaccessible Island, and in shaded places on Nightingale Island. On Gough Island 
Wace found it in thick Phylica forest, occasionally growing epiphytically near the ground. 

Tristan: Bonomi 25 (NBG [herb. Mus. Austro-Afric. 1114], K); Carmichael s.n. (K -holotype of 
variety, BM); Christophersen 457 (O, BM, K), 644 (O), 1243 (O); Dickson 10 (BM), 153 (BM); Keytel 
1829 (K), 1830 (K); Mejland 614 (O, BM, K); Moseley s.n. (K); Wace T.39 (BM). Inaccessible: 
Christophersen 2328 (O, BM), 2380 (O), 2526 (O, K), 2557 (O); Dickson 125 (BM). Nightingale: 
Christophersen 2186 (O, BM), 2220b (O). Gough: Wace 140 (BM), 156 (BM). 


ASPLENIACEAE 
ASPLENIOIDEAE 


Asplenium alvarezense R. N. R. Brown 
Endemic. First collected on Gough Island by R. N. R. Brown in April 1904, and described 
the following year (Brown, 1905: 247). It was not found on Tristan da Cunha until 1937-38, 
when Christophersen and Mejland discovered it there at a number of localities, it obviously 
having been overlooked by all previous visitors. On Gough Island it is dominant in the 
ground flora beneath Histiopteris incisa var. carmichaeliana over large areas. It has also been 
found on Inaccessible Island. 

Tristan: Christophersen 548 (O, BM, K), 1017 (O); Mejland 1389 (O, BM), 1390 (O, BM, K), 1550 
(O, BM, K), 1563 (O, BM, K). Inaccessible: Christophersen 2342 (O). Gough: Brown s.n. (E -holotype); 
Wace | (BM), 39 (BM), 62 (BM). 


Asplenium erectum Bory 

A insulare Carmich.; A. marinum sensu Thouars, non L. 

Native. First collected on Tristan by Aubert Du Petit-Thouars in January 1793. It has also 
been found on Inaccessible, Nightingale and Gough Islands, and occurs on wet cliffs, in 
moist gullies, and on damp shaded rocks. Although Christensen (1940: 13) found variation 
between examples from Tristan da Cunha, he was unable to discover any good character by 
which they could be distinguished from A. erectum of South Africa, although he did describe 
the examples with most deviation as a variety, var. aequibasis (see below). 

Tristan: Carmichael s.n. (K, BM—both syntypes of A. insulare); Mejland 1406 (O); Thouars s.n. (P, 
BM). Inaccessible: Christophersen 2344 (O); 2389 (O, BM), 2478 (L), 2580 (O), 2584 (O, B, K); 
Dickson 133 (BM, AAS); Moseley s.n. (K). Nightingale: Christophersen 2216 (O), 2227 (O, K); Dickson 
139 (BM); Moseley s.n. (K); Stableford 12 (K). Gough: McKinnon s.n. (K); Wace 56 (BM), 131 (BM), 
157 (BM). 

The following specimens are recognised by Christensen (1914: 14) as var. aequibasis 
Chr. 

Tristan: Mejland 1140 (O, BM, K —all intermediate between this and the type variety). Nightingale: 
Christophersen 2181 (O -holotype). 


Asplenium monanthes L. 
Native. First collected on Tristan da Cunha by Moseley in 1873 and since found several 
times mostly amongst moss in gullies and rock crevices. 

Tristan: Carmichael s.n. (BM, E [ex herb. Menzies]); Christophersen 233 (O, BM), 421 (O), 519 (O, 
BM, K), 550 (O), 1240 (O, BM, K); Dickson 110 (BM), 168 (BM); Mejland 823 (O), 1139 (O, BM, K), 
1392 (O); Moseley s.n. (BM); Wace T.46. 


Asplenium obtusatum var. crassum (Thouars) C.Chr. 
A. crassum Thouars 
Endemic variety. First found on Tristan da Cunha by Aubert Du Petit-Thouars in 1793 and 


* Du Petit-Thouars arrived in Réunion in 1795, where he spent some time gathering vascular plants, so confusion 
may have arisen during, or subsequent to, the time when he came to sort and label his collections. 


406 E. W. GROVES 


described as Asplenium crassum (Aubert Du Petit-Thouars, 1808: 71). It was subsequently 
reduced by Christensen in Christophersen (1937: 12) to a variety of A. obtusatum G. Forster, 
the type variety ofwhich is found in Australia and New Zealand. The var. crassum differs in 
having much more densely serrated pinnae, which are more distinctly unequal-sided at the 
base (i.e. truncate above and more or less excised below) and by its more numerous and 
larger, ovate-acuminate scales on the rachis and underside (see also Christensen, 1940: 12). 
On Tristan da Cunha the variety often occurs on bare rock or in shaded rock crevices, and on 
Gough Island it has been found in caves where they may be only a small amount of light. 

Tristan: Carmichael s.n. (BM, K, E [ex Wernerian volume]); Christophersen 424 (O, BM, K), 446 
(O); Dickson 45a (BM); Dyer 3561 (K, NBG); Keytel 1791 (K, NBG), 1832 (K, NBG); Mejland 145 (O, 
BM, K), 1160 (O), 1307 (O, BM, K), 1676 (O); Moseley s.n. (BM, K, E); Rogers s.n. (K); Siggeson s.n. 
(O); Thouars s.n. (P -holotype); Wace T.21 (BM). Inaccessible: Christophersen 2520 (O, BM); Moseley 
s.n. (BM, K); Nightingale: Christophersen 2097 (O); Dickson 134 (BM), 137 (BM); Glass s.n. (BM); 
Moseley s.n. (BM, K); Rogers s.n. (K); Stableford 8 (K), 14 (K), s.n. (K). Middle: Christophersen 2022 
(O, BM, K). Stoltenhoff: Christophersen 2152 (O, K). Gough: Brown s.n. (E); Fleming 20 (E), 29 (E), 38 
(E); MacKinnon s.n. (K); Swain s.n. (K); Wace 9 (BM), 40 (BM). 


Asplenium platybasis var. subnudum C.Chr. 

Endemic variety. Only a single gathering has been made of this interesting fern in Phylica 

forests on Inaccessible Island in 1938. The type of the species is known only from St Helena. 
Inaccessible: Christophersen 2501 (O -holotype, BM, K). 


ATHYRIOIDEAE 


Athyrium medium (Carmich.) Moore 
Aspidium medium Carmich.; Asplenium medium (Carmich.) Hook. 


Endemic. First collected on Tristan da Cunha by Carmichael, 1816-17. It grows on wet rock 
in gullies and cliffs and occasionally be streamsides. 

Tristan: Carmichael s.n. (BM, K-holotype); Christophersen 515 (O, BM), 1269 (O); Dickson 177 
(BM); Mejland 1370 (O, BM, K), 1553 (O). Inaccessible: Christophersen 2467 (O), 2468 (O, BM, K), 
2503a (O), 2522 (O). 


DRYOPTERIDOIDEAE 


Polystichum mohrioides (Bory) Presl 

Aspidium mohrioides Bory 
Native. A. Gepp identified this fern among the cryptogams collected by G. Wilkins, 
naturalist to the Shackleton-Rowett Expedition, 1921-22, when the Quest visited the group 
on its homeward voyage. Wilkins had found it on both Tristan da Cunha (vide Gepp’s list of 
determinations of the Shackleton-Rowett cryptogams in BM) and also on Gough Island 
(Wilkins, 1925: 70). It is a southern species occurring in South Georgia, Amsterdam Island, 
Marion Island, the Falkland Islands and in the Andes of South America. Modern con- 
firmation of its presence on Tristan da Cunha group is desirable. 

Tristan: Wilkins 38 (BM)*. Gough: Wilkins 51 pro parte (BM)*. 


Ctenitis aquilina (Thouars) Pichi-Serm. 


Polypodium aquilinum Thouars; Nephrodium aquilinum (Thouars) Hemsley; Dryopteris aquilina 
(Thouar) C. chr. 


Endemic. First collected on Tristan da Cunha by Aubert Du Petit-Thouars in January 1793. 
On Tristan da Cunha it is acommon understorey associate of the Phylica arborea forest, and 
on Inaccessible Island it occurs beneath the tree fern canopy. On Nightingale Island it has 
been found in the tussock grass community, and on Gough Island in sheltered gullies and 
beneath Phylica arborea. 

Tristan: Barkley in HMS Challenger s.n. (BM); Carmichael s.n. (E[herb. E. H. Layard and ex. 
When searching the herbarium folders at the BM in March 1979, I was unable to locate the original voucher 


specimens. As some of Wilkins specimens were only scraps it is possible that his material was not kept after 
identification. 


* 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 407 


Wernerian volume]), s.n. (BM); Christophersen 23 (O, BM), 423 (O, BM, K), 1061 (O), 1093 (O); Dyer 
3554 (PRE, NBG, BM); Fleming 52 (E), 72 (E), 90 (E), s.n. (E); Keytel 1828 (NBG, BM); Macgillivray 
s.n. (BM); Mejland 1575 (O); Moseley s.n. (BM, E); Thouars s.n. (P -holotype); Wace T.42 (BM), T.44 
(BM). Inaccessible: Christophersen 2545 (O); Moseley s.n. (BM). Nightingale: Christophersen 2003 
(BM, K), 2217 (O), 2232 (O); Fleming 57 (E). Gough: Fleming 18 (E); van der Merwe 6 (PRE), 53 bis 
(PRE); Wace 12 (BM), 133 (BM); Wilkins 88 (BM), 89 (BM) and 90 (BM). 


Dryopteris wallichiana (Sprengel) Hylander 

Aspidium paleaceum Sw.; A. parallelogramma Kunze; Dryopteris parallelogramma (Kunze) Alston; 
D. paleacea (Sw.) Hand-Haz 
Native. First collected by Wace in 1956 from boggy places away from shade but with some 
shelter. It has fronds up to 1.5 m in height. This species is pan-tropical and is related to the 
European Dryopteris affinis (Lowe) Fraser-Jenkins (D. borreri Newman ex von Tavel: D. 
pseudomas (Wollaston Holub & Pouzer), for which the name paleacea has sometimes been 
mistakenly used. 

Gough: MacKinnon s.n. (K); Wace 68 (BM). 


ELAPHOGLOSSOIDEAE 


Elaphoglossum hybridum (Bory) Moore 

Acrostichum ciliare Thouars: A. hybridum Bory 
Native. First collected on Tristan da Cunha by Aubert Du Petit-Thouars in 1793 and 
published as Acrostichum ciliare. The Tristan-Gough taxon has since been equated by 
Christensen (1940: 20) with the typical form of Elaphoglossum hybridum from Réunion and 
south Brazil. On Tristan it grows sporadically on moss-covered rock or in gullies in the 
mixed Phylica/Blechnum palmiforme association. 

Tristan: Carmichael s.n. (E [herb. Menzies and ex Wernerian volume]); s.n. (BM, K); Christophersen 
5501 (O), 690 (O, BM, K); Dickson 23 (BM, AAS); Mejland 617 (O, BM, K), 812 (O), 813 (O), 1148 (O, 
BM), 1617 (O); Thouars s.n. (P -holotype of A.ciliare); Wace T.32 (BM). Inaccessible: Christophersen 
2362 (O, BM, K). Gough: Wace 160 (BM). 


Elaphoglossum insulare C.Chr. 
Endemic. First collected at Hottentot Gulch on Tristan da Cunha by Mejland in January 
1938. It was found again by Dickson in 1962 growing on moss-covered rock at the same 
locality. 

Tristan: Dickson 167 (BM, AAS); Mejland 1386 (O, BM, K), 1618 (O -holotype, BM). 


Elaphoglossum laurifolium (Thouars) Moore 

Acrostichum laurifolium Thouars; A. conforme Carmich., non Swartz 
Endemic. First collected on Tristan da Cunha by Aubert Du Petit-Thouars in 1793. On Gough Island 
Wace in 1955 found it to be abundant amongst the Empetrum and tree-fern scrub at 355 m, and 
generally common on steep banks at lower altitudes. 

Tristan: Carmichael s.n. (K -holotype of A. conforme, E [herb. Menzies]); Christophersen 449 (O, 
BM, K); Dickson 9 (BM); Dyer 3557 (PRE, NBG); Keytel 1831 (NBG), 1846 (NBG); Mejland 342 (O), 
615 (O, BM, K); Thouars s.n. (P -holotype of A. /aurifolium), Wace T.35 (BM). Inaccessible: 
Christophersen 2548 (O), 2582 (O). Nightingale: Christophersen 2092 (O, BM, K), 2115 (O); Crosbie 
s.n. (E); Stableford 11 (K). Gough: Brown s.n. (K, E): Fleming | (E); van der Merwe 77 (PRE); Wace 37 
(BM), 114 (BM). 


Elaphoglossum obtusatum (Carmich.) C.Chr. 

Acrostichum obtusatum Carmich. 
Endemic. First collected on Tristan da Cunha in 1816-17 by Carmichael. It occurs on 
moss-covered rocks in gullies and under Phylica arborea. 

Tristan: Carmichael s.n. (BM, K -holotype, E [ex herb. Menzies]); Christophersen 538 (O, BM); 
Dickson 184 (BM); Mejland 811 (O, K), 1152 (O, BM), 1168 (O, BM), 1616 (O, K). Inaccessible: 
Christophersen 23 16b (O), 2590 (O). 


Elaphoglossum succisaefolium (Thouars) Moore 


Acrostichum succisaefolium Thouars 
Endemic. First collected on Tristan da Cunha by Aubert Du Petit-Thouars in 1793. On 


408 E. W. GROVES 


Tristan da Cunha it is common in the Phylica/Empetrum scrub, especially on steep banks in 
association with Blechnum palmiforme and Blechnum penna-marina at altitudes mostly 
between 30-300 m. On Inaccessible Island it grows amongst the tree ferns, and on 
Nightingale Island it has been found on ledges and cliffs. On Gough Island it occurs 
frequently on steep peaty banks and on decaying trunks of old tree fern. It is by far the 
commonest species of Elaphoglossum occurring on the islands. 

Tristan: Barkly in HMS Challenger s.n. (BM); Carmichael s.n. (BM, K, E [ex Wernerian volume)]); 
Christophersen 208 (O); Dickson 95 (BM); Dyer 3556 (PRE, NBG, K); Keytel 1843 (NBG, K), 1844 
(NBG, K), s.n. (PRE [herb. Marloth 4702]); Macgillivray 328 (K), s.n. (BM); Mejland 156 (O, BM, K), 
343 (O, BM, K), 616 (O), 821 (O), 1409 (O), 1635 (O); Milne s.n. (K); Moseley s.n. (K, BM); Rogers s.n. 
(K); Thouars s.n. (P -holotype, K); Wace T.3a (BM), T.31 (BM). Inaccessible: Christophersen 2307, 
(O), 2361a (O), 2556 (O, BM); Moseley s.n. (K). Nightingale: Christophersen 2000 (O, BM, K), 2042 
(O, BM), 2091 (O, K): Crosbie s.n. (E); Moseley s.n. (K); Stableford s.n. (K). Gough: van der Merwe 19 
(PRE); Wace 38 (BM), 59 (BM), 70 (BM), 115 (BM). 


DAVALLIACEAE 
DAVALLIOIDEAE 


Rumohra adiantiformis (G. Forster) Ching 

Polypodium adiantiforme G. Forster; Polystichum adiantiforme (G. Forster)J.Sm.; Aspidium 
coriaceum Sw.; Polypodium calyptratum Thouars 
Native. First collected on Tristan da Cunha by Aubert Du Petit-Thouars in 1793 and 
published as a new species, Polypodium calyptratum (Aubert Du Petit-Thouars, 1808: 33). It 
has been found several times on the island since its first discovery, mostly on cliffs, and also 
on Inaccessible and Gough Islands. On the latter Wace found that in 1966 it was plentiful on 
dry peat banks particularly where drainage was good. Christensen (1940: 18) comments that 
the species exhibits much variation throughout its geographical range and that it is possible 
that the Tristan-Gough taxon, which is closely related to some forms from Argentina and 
South Africa, may deserve varietal rank. 

Tristan: Bonomi s.n. (NBG [herb. Mus.Austro-Afric. 1112]); Carmichael s.n. (E [ex herb. E. H. 
Layard]); Christophersen 17 (O, BM, K); Dickson 147 (BM, AAS), 179 (BM); Dyer 3562 (PRE, NBG); 
Fleming 39 (E); Glass s.n. (BM); Mejland 153 (O), 1410 (O, BM, K); Moseley s.n. (BM, E); Siggeson s.n. 
(O); Thouars s.n. (P -holotype of Polypodium calpytratum). Inaccessible: Christophersen 2571 (QO); 
Moseley s.n. (K, P). Gough: Fleming 11 (E), 23 (E); van der Merwe 52 (PRE); Wace 36 (BM), 90 (BM). 


BLECHNACEAE 


Blechnum australe L. 

Native. First collected on Tristan da Cunha by Aubert Du Petit-Thouars in January 1793 
and published in his list (Aubert Du Petit-Thouars, 1808: 33). It grows commonly in gullies 
and rock crevices and sometimes on cliffs beneath Phylica or on boulder scree. 

Tristan: Bonomi 26 (NBG [herb. Mus. Austro-Afric. 1115]), K, 28 (PRE, E); Carmichael s.n. (K, E 
[ex herb. E. H. Layard and ex Wernerian volume]); Christophersen 428 (O, BM, K), 1242 (O, BM); 
Crosbie s.n. (E); Dickson 155 (BM, AAS); Dyer 3522 (PRE, K); Dyer (coll. by an islander) s.n. (PRE, 
NBG, K); Keytel 1823 (NBG, K), s.n. (PRE [herb. Marloth 4718]); Macgillivray 331 (K), 332 (K); 
Mejland 111 (O), 158 (O), 816 (O), 1164 (O, BM, K), 1306 (O, K); Milne s.n. (K); Moseley s.n. (K, E); 
Rogers s.n. (K); Siggeson s.n. (O); Thouars s.n. (P); Wace T.25 (BM). Inaccessible: Christophersen 
2308 (O), 2577 (O), 2600 (O, BM); Moseley s.n. (K, E). Gough: Wace 94 (BM), 100 (BM). 

Blechnum palmiforme (Thouars) C.Chr. TRISTAN TREE FERN 

Pteris palmaeformis Thouars; Lomaria robusta Carmich. L. boryana sensu Hemsley, non Willd. 
Endemic. First collected on Tristan da Cunha by Aubert Du Petit-Thouars in 1793. It is 
distinguished from the South African species Blechnum tabulare (Thunb.) Kuhn, with which 
it has by some authors been united, by having the pinnae narrowed towards the base of the 
frond (Christensen, 1940: 11). (See Figs 11, 12 & 33). 

Tristan: Carmichael s.n. (BM -holotype of L. robusta, E), s.n. (E[ex Wernerian volume and ex Herb. 
Menzies]); Christophersen 45 (O, BM), 642 (O, BM, K); Crosbie s.n. (E); Dickson 154 (BM, AAS), 163 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 409 


ad 


we 


Fig. 33 Gough Island near che “idee ae with the ferns cite atc rsom and Histiopteris 
incisa var. carmichaeliana and the island tree, Phylica arborea, forming the foreground. 8 June 
1927. From an original watercolour by Sir Alister Hardy. 


(BM), 174 (BM); Dyer 3555 (PRE, NBG); Fleming 78 (E); Mejland 157 (O); 1014 (O, K), 1437 (O), 
1438 (O, BM), 1439 (O, BM), 1440 (O, BM, K), 1441 (O); Thouars s.n. (P -holotype of P. 
palmaeformis); Wace T.40 (BM). Inaccessible: Christophersen 2440 (O, BM), 2455 (O); 2559 (O); 
Stableford 121 (K). Nightingale: Christophersen 2212 (O, BM, K). Gough: Brown s.n. (E): MacKinnon 
s.n. (K); van der Merwe 40 (PRE); Wace 55 (BM), 67 (BM), 153 (BM), 154 (BM). 


Blechnum penna-marina (Poiret) Kuhn 

Acrostichum polytrichoides Thouars (text) [‘A. polypodioides’ below plate]; Lomaria antarctica 

Carmich.; L. alpina sensu Hemsley, non (R.Br.) Spreng. 
Native. First collected on Tristan da Cunha by Aubert Du Petit-Thouars in January 1793. 
Here it has now become abundant on the steeper slopes, where it often forms a dense pure 
community; it also occurs up to 650 m in scrub associated with Empetrum rubrum and 
Phylica arborea. Examples cytologically investigated by Manton & Vida (1968: 372-3) were 
shown to differ from New Zealand plants of the same species that had long been under 
cultivation in Leeds University greenhouses. The Tristan material gave 2n=66; that from 
New Zealand 2n=68. In addition, the morphological differences between the Tristan and 
South American plants in comparison with those from Australasia indicate separation at 
subspecific level (A. C. Jermy, pers. comm.) 

Tristan: Bonomi 26a (NBG), 28 (K); Carmichael s.n. (BM -holotype of L. antarctic s.n. (E [ex herb. 
Menzies]); Christophersen 4b (O), 209 (O), 471 (O), 1062 (O); Dickson 13 (BM); Dyer 3521 (PRE, K, 
NBG); Fleming 42 (E), 56 (E), 87 (E); Glass s.n. (BM); Keytel 1824 (NBG, K, BM). s.n. (PRE [herb 
Marloth 4718a]); Macgillivray 333 (K), s.n. (BM); Mejland 112 (O), 640 (O, BM, K); Milne s.n. (K); 
Moseley s.n. (BM); Siggeson s.n. (O, BM); Stableford 67 (BM), 81 (BM); Thouars s.n. (P -holotype of A. 
polytrichoides, BM); Wace T.24 (BM), si 34 (BM), T.66 (BM). Inaccessible: Christophersen 2553 (O). 
Nightingale: Christophersen 2037 (O); Stableford 13 (K), s.n. (K). Gough: Brown s.n. (K); MacMillan 
s.n. (PRE); van der Merwe 4 (PRE); Wace 13 (BM), 80 (BM), 106 (BM). 


410 E. W. GROVES 


AZOLLACEAE 


Azolla filiculoides Lam. 

Native. Found only on Inaccessible Island, where two collections were made in 1938—one 
in a swamp at West Point, and the other in a brook on the beach at Blenden Hall. Dickson 
noted in 1962 that it was abundant on ‘wet ground under Spartina, and especially on open 
water’ at these two localities (Wace & Dickson 1965: 317 & photograph: Fig. 30). 
Christensen (1940: 23) considers that this widely distributed American species is probably of 
recent introduction. 

Inaccessible: Christophersen 2446 (O), 2615 (Q, BM, K). 


Comments on the flora 


Altogether 212 flowering plants and ferns are known from the islands of the Tristan da 
Cunha group. This figure comprises both native and introduced taxa at specific level or 
below. The native taxa are further divided into those considered endemic (i.e. not recorded 
outside of the islands) and those considered native in the strict sense (i.e. present on the 
islands for a very long time and with a wider distribution beyond the group): 


Angiosperms Total taxa 
Endemic 34 
Native (s.str.) 24 
Introduced 119 
Total 177 


Pteridophytes 


Endemic 20 
Native (s.str.) \I5) 
Introduced 0 
Total 35 


A further breakdown of the above figures to show the totals for each island is as follows: 


Tristan Inaccessible Nightingale Gough 

Area (in sq. km) 86 W 4 57 
Altitude (in m) 2060 C.F c. 300 910 
Spermatophyte taxa 

Endemic Di 20 14 21 

Native (s.str.) 20 17 7 15 

Introduced 115 22 5 17 
Pteridophyte taxa 

Endemic 19 I 7/ 12 15 

Native (s.str.) 14 10 4 12 
Total number of 
vascular plant taxa 195 86 42 80 


As a result of critical study by Carl Christensen, A. H. G. Alston and A. F. Tyron, the 
pteridophytes are probably the most completely known and best understood taxonomically 
of all the major plant groups on the islands. Although it has been possible to correct hitherto 
uncertain names for some ferns it is some measure of the extent of the work that the number 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 41] 


of taxa for this group of plants has only been increased by two since the 1965 list of Wace & 
Dickson. The two additions have come about as a result of generic revisions rather than by 
new discoveries in the field. Although it is probably unlikely that many new species of fern 
remain to be discovered on the islands, there is nevertheless much culture experimental work 
that still needs to be done on some of the species already known to occur in the Tristan da 
Cunha group. 

It is not the purpose here to repeat the analytical and biological data on the flora given in 
Wace & Dickson (1965), and the reader is referred to that paper for the valuable 
information it contains. However, certain tables are presented here in a similar form so that a 
ready comparison may be made between the status of the flora then with that as it is known 
today. It has been possible during research for this present paper to examine all the hitherto 
unpublished collections of the Tristan-Gough vascular plants in the major herbaria which 
were not readily available to the authors just mentioned. It has also been possible to make 
taxonomic evalution on a number of species so as to bring their nomenclature in line with 
recent revisions. 


Relationship of the flora of the Tristan da Cunha group of islands 

By far the greatest number of native and endemic vascular plants of these islands have either 
a South American or a south circumpolar distribution, or are closely allied to species that 
have such a range. Although the islands are closer to southern Africa (see Fig. 2), the African 
element (i.e. related to species on that mainland) is with regard to the flowering plants quite 
small. Indeed, there is no native phanerogam present on the Tristan da Cunha group that is 
otherwise entirely confined to Africa. With the ferns, however, whilst 30 out of the islands’ 
35 taxa have close relatives in the southern part of South America, or are known from other 
Antarctic islands, the African connection in this group of plants is much stronger, with two- 


Table 1 Spermatophytes of the Tristan da Cunha group of islands. Figures in brackets indicate the 
number of endemics. 


Whole Group _ Tristan Inaccessible Nightingale Gough 


Native (s.lat.) 


Number of families 19 16 15 1] 16 
Number of genera a2 28 D'S 16 25 
Number of taxa at species 

level or below 58 (34) 47 (27) 37 (20) 21 (14) 36 (21) 


Number of taxa at species 
level or below recorded on 


one island only 18 (11) 9 (6) 2 (0) 1 (1) 6 (4) 
Introduced 

Number of families 33 32 12 3 3 

Number of genera 83 80 20 5 10 

Number of taxa at 

species level or below 119 115 22 5 UF 


Total number of taxa at 

species level or below 

of both native (s.lat.) 

and introduced plants 177 162 59 26 53 


Number of visits on 
which spermatophytes have 
been collected 28 7 7 7 8 


412 E. W. GROVES 


Table 2 Pteridophytes of the Tristan da Cunha group of islands. Figures in brackets indicate the number 
of endemics. 


Whole Group Tristan Inaccessible ‘Nightingale Gough 
Native (s.lat.) 

Number of families 15 14 14 11 12 
Number of genera 20 19 18 13 7 
Number of taxa at 
species level or 
below 35 (20) 33 (19) 27 (7) 16 (12) 27 (15) 
Number of taxa at 
species level or below 


recorded on one island only 7 (3) 4 (2) 2 (1) 0 (0) 1 (0) 
Number of visits on 
which pteridophytes 
have been collected 22 15 4 7 8 


Table 3 Native (s.str.) & endemic vascular plants known only from a single island in the Tristan da 
Cunha group. (N = native; E = endemic) 


Species known only from Tristan da Cunha: 


Nertera depressa forma fimbriata (E) Agrostis wacei (E) 

Calystegia soldanella (N) Agrostis sp. (Phalaris cespitosa) (E) 
Carex thouarsii var. curvata (E) Asplenium erectum var. aequibasis (E) 
Scirpus chlorostachyus (N) Asplenium monanthes (N) 

Scirpus verruculosus (N) Elaphoglossum insulare (E) 

Agrostis crinum-ursi (E) Hymenophyllum tunbrigense (N) 


Agrostis holgateana (E) 


Species known only from Inaccessible Island: 
Peperomia berteroana (N) Asplenium plastybasis var. subnudum (E) 
Calystegia tuguriorum (N) Azolla filiculoides (N) 


Species known only from Nightingale Island: 
Cotula moseleyi (E) 


Species known only from Gough Island: 
Sophora microphylla (N) Dryopteris wallichiana (N) 
Cotula goughensis (E) 
Tetronicium magellanica (N) 
Agrostis goughensis (E) 
Deschampsia robusta (E) 
Deschampsia wacei (E) 


(N.B. With more intensive collecting some of these species will undoubtedly be found on other islands 
of the group) 


thirds of the taxa affiliated to that continent. Few of the pteridophytes have a range extending 
northwards beyond the equator into the northern hemisphere tropics, and only one, 
Hymenophyllum tunbrigense, has a distribution which includes areas within the north 
temperate zone. 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 413 


Distribution of vascular plants within the islands 

Tables 1-3 give a comparison of the numbers of families, genera and specific taxa occurring 
on each island for both flowering plants and ferns. Tristan da Cunha (the only inhabited 
island) has been visited and botanical collections made on it on a good number of occasions, 
but Inaccessible and Nightingale on the other hand, have received only few visits from 
botanical collectors. 


Dispersal mechanisms of the native species 

A detailed discussion on each of the three ways (by wind, sea and birds) by which the native 
terrestrial flowering plants could have arrived on the Tristan da Cunha islands has already 
been given by Wace & Dickson (1965: 301-309). Suffice it to mention here those authors’ 
general conclusion that the evidence at present available suggests that external transport by 
birds of the plant diaspores (i.e. by hooked adhesion to feathers or down or by viscid 
adhesion to feet) ‘has been the most important agency by which terrestrial plants have 
arrived on the islands; that sea flotation has been of some importance; and that air flotation 
has only been of importance in the dispersal of cryptogams’. 


Spread of alien flowering plants 
Whilst the relationship of the native species (s.lat.) of any remote oceanic island to the flora 
of adjacent land masses is always of importance, the arrival of alien species has become of 


Table 4 Number of alien species found by the different plant collectors, arranged in chronological order 
of collection. The numbers in brackets give the progressive total at each date. 


Date of visit Collector Tristan Inaccessible Nightingale Gough 
1793 Aubert Du 

Petit-Thouars 3 = = = 
1816-17 Carmichael 7 {(>))) - - - 
1852 Macgillivray & 

Milne 10 (15) = = = 
1873 Moseley 3 (18) 5 = = 
prior to 1884 vide Hemsley 1 (19) - - — 
1904 Bonomi 12 (31) - - - 
1904 Brown - - - 4 
1908-09 Keytel 9 (40) - - = 
prior to 1910 referred to in 

Mrs Burrow’sbook 4 (44) = = = 
1922-25 Rev. and Mrs Rogers 1 (45) - - - 
1927 Discovery (W.S.) 

Expedition - - - 1 (5) 
1934 Siggeson 3 (48) = = = 
Feb-Mar 1937 Dyer 5 (53) = = = 
1937-38 Christophersen 

& Mejland 26 (79) 11 (16) 3 - 
1953-54 Stableford 16 (95) = 1 (4) = 
1955-56 Stableford 3 (98) = = a 
1955-57 Wace 2 (100) - - TG) 
1956-57 van der 

Merwe = = - 4 (16) 
1962 Dickson 6 (106) 4 (20) = = 
1968 Wace WCU) DO) = = 
1972 Fleming - = = 1 (17) 
1976 Wace 2 (115) - 1 (5) - 


414 E. W. GROVES 


increasing interest to biogeographers and students of plant dispersal. From the date of the 
first botanical investigation of the islands of the Tristan group in 1793 (when, along with 
native species, the earliest aliens Chenopodium album, Lactuca scariola and Raphanus 
sativus, were first noted) until 1899, 19 aliens had been found on Tristan and five on 
Inaccessible Island (see Table 4). From 1900 to 1929, 26 more were found on Tristan 
(making a total at the end of that period of 45), five on Inaccessible Island (no increase) and 
five on Gough Island. During the years 1930 to 1939 another 34 aliens were recorded on 
Tristan da Cunha (bringing the total to 79); 11 more on Inaccessible Island (bringing the total 
to 16); and three were found on Nightingale. No collecting was done on the islands during 
the decade 1940-1949, but from 1950 until 1968, due to more intensive and systematic 
collecting, the aliens for Tristan were increased by another 34, bringing the total there to 
113; six more had been added from Inaccessible Island, bringing that island’s total to 22; one 
more for Nightingale making four in all; and another 11 to those previously found on Gough, 
bringing the total to 16. The most recent aliens recorded were the two grasses Festuca 
arundinacea subsp. mediterranea and Pennisetum clandestinum found on Tristan in 1976 
by Wace, bringing the total for that island to 115. Phormium tenax was noted on Nightingale 
bringing the total of aliens to five; and Cerastium fontanum subsp. triviale from Gough 
brings that islands total to 17. These numbers will undoubtedly be increased when all the 
material gathered during the 1968 and 1976 visits are distributed and become available for 
study (see p. 353). There have been no alien ferns recorded from the islands. 

With perhaps the exception of Chenopodium album, C. murale, Lactuca scariola, 
Raphanus sativus, and Rumex crispus (all from Tristan da Cunha) the introductions are 
likely to be still present on the group. Most of these species seem to have originated from the 
Palaearctic region and to have arrived on the islands via the Cape Province of South Africa. 
Their numbers markedly increased during the period in which grazing animals, such as 
cattle, sheep and goats, were introduced. It will be of considerable interest to botanists to 
observe what further colonizers may be added to the alien flora of the Tristan da Cunha 
group of the islands in the years to come. 


Acknowledgements 


I am indebted to the Directors of the following Institutes for facilities for working in their 
herbaria or for the loan of specimens from the collections under their charge: Botanical 
Museum, University of Oslo (O); Botany School, University of Cambridge (CGE); Botanical 
Section, British Antarctic Survey (AAS) (then at Birmingham now at Cambridge); Compton 
Herbarium, National Botanic Gardens, Kirstenbosch, Cape Town (NBG); Laboratoire de 
Phanérogamie, Muséum National d’Historie Naturelle, Paris (P); National Herbarium, 
Botanical Research Institute, Pretoria (PRE); Royal Botanic Garden, Edinburgh (E), and the 
Royal Botanic Gardens, Kew (K). 

For the illustrations accompanying this paper I wish to acknowledge my thanks to Sir 
Alister Hardy (Oxford) for permission to reproduce two of his original watercolours, to Dr 
Erling Christophersen (Oslo) for permission to reproduce Fig. 2, to Sir Hugh Elliott and the 
Committee of the International Union for Conservation of Nature & Natural Resources for 
the use of the other maps which were previously published in their Monograph No. 6, to Dr 
Christophersen, Dr H Heine (Museum National d’Histoire Naturelle, Paris) and to the 
librarians of the Royal Botanic Gardens, Kew, and the Hunt Institute for Botanical 
Documentation, Pittsburgh, U.S.A., for the use of some portraits, and to Dr Nigel Wace 
(Canberra, Australia) and Mr George Edwards (Eastleigh, Hampshire) for kindly providing 
me with the vegetational and topographical photographs. The drawings of grasses were 
kindly prepared by Ann Davies, and the Bentham-Moxon Trust is thanked for financial 
assistance. 

I am also grateful to the following people for help in various ways: Dr C. E. Hubbard, Kew 
(who regrettably died in May 1980), for providing accounts of the new taxa in the 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 415 


Gramineae, Messrs J. A. Crabbe and A. C. Jermy (British Museum [Natural History]) and 
Professor R. E. Holttum (Kew) and Dr B. S. Parris (Cambridge) for assistance with problems 
relating to the pteridophytes, Miss S. Hooper (Kew) for advice in connection with 
Cyperaceae, Dr A. Lourteig of the Laboratoire de Phanérogamie, Museum National 
d’Histoire Naturelle, Paris, for aid in my search there for Aubert Du Petit-Thouars types, 
and Mr J. R. Laundon for help in preparing the manuscript for publication. 

Finally I wish to record my special thanks to Dr Stanley Greene, lately Head of the 
Botanical Section, British Antarctic Survey (now at the Institute of Terrestrial Ecology, 
Edinburgh), for his continued encouragement and suggestions during helpful discussions, 
and also to Mrs Dorothy Greene, for her invaluable assistance in processing all the record 
data used in this paper. 


References 


Aellen, P. 1928. Die Chenopodium-Arten des siidlichen und mittleren Africa. Reprium Spec. nov. 
Regni veg. 24 : 337-347. 

Agardh, J. G. 1839. Recensio Specierum Generis Pteridis. Lund. 

Aubert Du Petit-Thouars, A. 1808*. Esquisse de la Flore de I'Isle de Tristan d’Acunga. Paris. 
{Reprinted in his Melanges de Botanique et des Voyages 25-46 and tabs I-XII (181 1)]. 

Baird, D. E. 1965. The effects of the eruption of 1961 on the fauna of Tristan da Cunha. Phil. Trans. R. 
Soc. B, 249 : 425-434. 

Baker, P. E., Gass, I. G., Harris, P. G. & LeMaitre, R. W. 1964. The volcanological report of the 
Royal Society Expedition to Tristan da Cunha 1962. Phil. Trans R. Soc. B, 249 : 257-434. 

Barnhart, J. H. 1965. Biographical Notes upon Botanists. 3 vols. Boston, Mass. 

Barrow, F. M. 1910. Three years in Tristan da Cunha. London. 

Bitter, G. 1911. Die Gattung Acaena. Vorstudien zu einer Monographie. Biblthca bot. 74: 1-336. 

Booy, D. M. 1957. Rock of Exile. A Narrative of Tristan da Cunha. London. 

Brander, J. 1940. Tristan da Cunha, 1506-1902. London. 

Brown, R. N. R. 1905. The botany of Gough Island. I. Phanerogams and ferns. J. Linn. Soc. (Bot.) 37 : 
238-250. 

—, Wright, C. H. & Darbishire, O. V. 1912. The botany of Gough Island [Scottish National Antarctic 
Expedition]. Rep. scient. Results Voy. S. Y. ‘Scotia’ 3 (Botany) : 33-46. 

Brummitt, R. K. & Groves E. W. 1981. The occurrence of Calystegia tuguriorum (Colvolvulaceae) in 
the Tristan da Cunha islands. Kew Bull. 35 (1) [in press]. 

Carmichael, D. 1819. Some account of the island of Tristan da Cunha and its natural productions. 
Trans. Linn. Soc. Lond. 12 : 483-513. 

Christensen, C. 1937. Pteridophyta. /n E. Christophersen, Plants of Tristan da Cunha. Scient. Results 
Norw. Antarct. Exped. 16 : 12-13. 

— 1940. The pteridophytes of Tristan da Cunha. Results Norw. scient. Exped. Tristan da Cunha 
6: 1-25. 

Christophersen, E. 1934. Plants of Gough Island (Diego Alvarez). Scient. Results Norw. Antarct. 
Exped. 13: |\-16. 

— 1937. Plants of Tristan da Cunha. Scient. Results Norw. Antarct. Exped. 16: 1-19. 

— 1940a. Tristan da Cunha, the Lonely Isle. London. 

— 19406. Ranunculus caroli, en ny art fra Tristan da Cunha. Acta phytogeogr. suec. 13 : 186-190. 

— 1944. New phanerogams from Tristan da Cunha. Results Norw. scient. Exped. Tristan da Cunha 
11: 1-15. 

— 1968. Flowering plants from Tristan da Cunha. Results Norw. scient. Exped. Tristan da Cunha 
55: 1-29. 

Crabbe, J. A., Jermy, A. C. & Mickel, J. T. 1975. A new generic sequence for the pteridophyte 
herbarium. Brit. Fern Gaz. 11: 141-162. 

Crawford, A. B. 1941.7 Went to Tristan. London. 

Dickson, J. H. 1965a. The biological report of the Royal Society expedition to Tristan da Cunha, 
1962. Part 1. General introduction. Phil. Trans. R. Soc. B, 249 : 259-271. 

— 19655. The effects of the eruption of 1961 on the vegetation of Tristan da Cunha. Phil. Trans. R. Soc. 
B, 249 : 403-424. 


* See Stafleu & Cowen (1976: 705) for publication details. 


416 E. W. GROVES 


Dunne, J. C. 1941. Volcanology of the Tristan da Cunha group. Results Norw. scient. Exped. Tristan 
da Cunha 2 : 1-145. 

Dyer, R. A. 1939. The flora of Tristan da Cunha. HMS Carlisle expedition, 1937. Bothalia 
3 : 589-607. 

Gilmour, J. S. L. & Tutin, T. G. 1933. A List of the More Important Collections in the University 
Herbarium, Cambridge. Cambridge. 

Good, R. D. O. 1927. The genus Empetrum L. J. Linn. Soc. (Bot.) 47 : 489-523. 

Green, L. G. 1973. When the Journey's Over. London. [Chapter 8 (pp. 84-102) refers to the islands. ] 

Gunther, E. R. 1928. Notes and Sketches Made During Two Years on the ‘Discovery’ Expedition 
1925-27. Oxford. 

Hafsten, U. 1951. A pollen-analytic investigation of two peat deposits from Tristan da Cunha. Results 
Norw. scient. Exped. Tristan da Cunha 22 : 1-42. 

— 1960. Quaternary history of vegetation in the south Atlantic islands. Proc. R. Soc. B, 152 : 516-529. 

— 1961. Pleistocene development of vegetation and climate in Tristan da Cunha and Gough Island. 
Arbok. Univ. Bergen Mat.-Natury. Ser. 1960 (20). 

Hardy, A. 1967. Great Waters. London. 

Hemsley, J. B. 1885. Report on the botany of the Bermudas and various other islands in the Atlantic 
and southern oceans. Scient. Results Voyage ‘Challenger’ (Bot.) 1: 113-185. 

Holdgate, M. 1958. Mountains in the Sea. The Story of the Gough Island Expedition. London. 

Holmgren, P. K. & Keuken, W. 1974. Index Herbariorum. Part 1. The Herbaria of the World. 6th ed. 
Utrecht. 

Holttum, R. E. 1974. Thelypteridaceae of Africa and adjacent islands. J/ S.A/fr. Bot. 40 : 123-168. 

Hooper, S. S. 1968. Cyperaceae from Tristan da Cunha. Results Norw. Scient. Exped. Tristan da 
Cunha 54: 1-9. 

Hubbard, C. E. 1954. Grasses. Harmondsworth. 

Knuth, R. 1912. Geraniaceae. In A. Engler, Das Pflanzenreich IV, 129 : 1-640. Leipzig. 

Linklater, E. 1972. The Voyage of the ‘Challenger’. London. 

Manton, I. & Vida, G. 1968. Cytology of the fern flora of Tristan da Cunha. Proc. R. Soc. B, 
170 : 361-379. 

Markham, K. R. & Godley, E. J. 1972. Chemotaxonomic studies in Sophora. 1. An evaluation of 
Sophora microphylla Ait. N.Z. JI Bot. 10 : 627-640. 

Moseley, H. N. 1874. Notes on the plants collected in the islands of the Tristan d’Acunha group. J. 
Linn. Soc. (Bot.) 14 : 377-384. 

— 1879. Notes by a Naturalist. An Account of Observations Made During the Voyage of HMS 
‘Challenger’. London. 

Munch, P. 1971. Crisis in Utopia. London. 

Palisot de Beauvois, A. M. F. J. 1804-07. Flore d’Oware et de Benin en Afrique. 2 vols. Paris. 

Parris, B. S. A revision of the genus Grammitis Sw. (Filicales: Grammitidaceae) in Australia. J. Linn 
Soc. (Bot.) 70 : 21-43. 

Phillips, E. P. 1913. A list of the phanerogams and ferns collected by Mr P. C. Keytel on the island of 
Tristan da Cunha, 1908-09. Ann. S. Afr. Mus. 9 : 96-103. 

Pichi Sermolli, R. E. G. 1957. Adumbratio Florae Aethiopicae. Pt. 5, Parkeriaceae, Adiantaceae, 
Vittarioaceae. Webbia 12 : 645-703. 

— & Bizzarri, M. P. 1978. The botanical collections (Pteridophyta and Spermatophyta) of the AMF 
Mares—G.R.S.T.S. Expedition to Patagonia, Tierra del Fuego and Antarctica. Webbia 32 : 
455-534. 

Rechinger, K. H. 1954. Monograph of the genus Rumex in Africa. Bot. Notiser Suppl. 3 (3): 1-114. 

Rogers, R. A. 1926. The Lonely Island. London. 

Rowan, M. K. 1952. The greater shearwater (Puffinus gravis) at its breeding ground. [bis 94 : 97-121. 

Skottsberg, C. 1946. Peperomia berteroana Mia. and P. tristanensis Christoph., an interesting case of 
disjunction. Acta Horti gothoburg. 16 : 251-288. 

Spry, W. J. J. 1877. The Cruise of HMS ‘Challenger’ Over Many Seas, Scenes in Many Lands. 2nd ed. 
London. 

Stafleu, F. A. & Cowan, R.S. 1976. Taxonomic Literature. A Selective Guide to Botanical Publications 
and Collections with Dates, Commentaries and Types. 2nd. ed. 1, A-G. [Regnum veg. 94] Utrecht. 
Swire, H. 1938. The Voyage of the ‘Challenger’. A Personal Narrative of the Historic Circumnavigation 

of the Globe in the Years 1872-1876. London. 2 vols. [pp. 82-91 of Vol. 1 refer to the islands.] 

Sykes, W. R. & Godley, E. J. 1968. Transoceanic dispersal in Sophora and other genera. Nature, Lond. 
218 : 495-496. 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 417 


Thomson, C. W. 1877. Voyage of the ‘Challenger’. The Atlantic. A. Preliminary Account of the General 
Results of the Exploring Voyage of H.M.S. ‘Challenger’ During the Year 1873 and the Early Part of 
the Year 1876. London. 2 vols. [pp. 151-187 of vol. 2 refer to the islands. ] 

Tryon, A. F. 1966. Origin of the fern flora of Tristan da Cunha. Brit. Fern Gaz. 9 : 269-276. 

— 1970. A monograph of the fern genus Eriosorus. Contr. Gray. Herb. Hary. 200 : 54-174. 

Verrill, G. E. 1895. On some birds and eggs collected by Mr George Comer at Gough Island, Kerguelen 
Island and the island of South Georgia. Trans. Conn. Acad. Arts. Sci. 9 : 132-136. 

Wace, N. M. 1960. Botany of the southern oceanic islands. Proc. R. Soc.B, 152 : 475-489. 

— 1961. The vegetation of Gough Island. Ecol. Monogr. 31 : 337-367. 

— 1967. Alien plants in the Tristan da Cunha islands. Jnt. Un. Consery. Nature Il, 9 : 46-60. 

— & Dickson, J. H. 1965. The terrestrial botany of the Tristan da Cunha islands. Phil. Trans. R. Soc. 
B, 249 : 273-360. 

— & Holdgate, M. W. 1958. The vegetation of Tristan da Cunha. J. Ecol. 46 : 593-620. 

— 1976. Man and nature in the Tristan Islands. nt. Un. Conserv. Nature Mongr. 6: 1-114. 

Wild, F. 1923. Shackleton’s Last Voyage. The Story of the ‘Quest’. 2nd ed. London. 

Wilkins, G. H. 1925. Gough Island. J. Bot., Lond. 63 : 65-70. 


Index to species 


Accepted names are given in roman and synonyms in italic; new names are in bold. An asterisk (*) 
denotes a figure. The letter ‘f indicates mention in a footnote. 


Acaena sanguisorbae 359 wacei 383, 384* 
sarmentosa 359 sp. 385 
stangii 359 Aira caryophyllea 385 


Acrostichum ciliare 407 Amauropelta bergiana var. tristanensis 404 
conforme 407 tomentosa 404 


hybridum 407 Anagallis arvensis 367 
laurifolium 407 Ancistrum sarmentosa 359 
obtusatum 407 Anthemis cotula 363 
polypodioides 409 Anthoxanthum odoratum 385 
polytrichoides 409 Apium australe 361 
succisaefolium 407 goughense 361 


Adiantum aethiopicum 402 
crenatum 402 
poiretti 402 
thalictroides 402f 
Agrostis carmichaelii 378 
castellana var. mixta 378 
crinum-ursi 379 
difficilis 382 
gigantea 379 
goughensis 379, 380* 
holgateana 379, 380* 
lachnantha 379 
magellanica subsp. laeviuscula 38 | 
media 382 
prostrata 382 
ramulosa 378 
simulans 382 
stolonifera subsp. stolonifera 382 
tenuis 382 
trachychlaena 383, 384* 
tropica 383 


Aspidium coriaceum 408 
medium 406 
mohrioides 406 
paleaceum 407 
parallelogramma 407 
Asplenium alvarezense 405 
crassum 405 
erectum 405 
erectum var. aequibasis 405 
filipendulaefolium 401 
insulare 405 
marinum 405 
medium 406 
monanthes 405 
obtusatum var. crassum 405 
platybasis var. subnudum 406 
Athyrium medium 406 
Atriplex plebeja 370 
Azolla filiculoides 410 


Bellis perennis 363 


418 E. W. GROVES 


Blechnum australe 408 
palmiforme 408 
penna-marina 409 
tabulare 408 

Brassica campestris 354 
juncea 354 
rapa 354 

Bromus catharticus 386 
unioloides 386 
willdenowii 386 


Calamagrostis deschampsiiformis 386, 387* 
Calendula pusilla 366 
Callitriche christensenii 360 
Calystegia sepium subsp. americana 368 
soldanella 368 
tuguriorum 368 
Cardamine glacialis 354 
propinqua 354 
Carex insularis 375 
thouarsii var. recurvata 375 
thouarsii var. thouarsii 375 
Centella asiatica 361 
Cerastium caespitosum 355 
caespitosum subsp. triviale 355 
fontanum subsp. triviale 355 
holosteoides 355 
triviale 355 
Cheilanthes viscosa 404 
Chenopodium album 370 
ambrosioides subsp. 
tomentosum 370 
murale 371 
tomentosum 370 
Chevreulia sarmentosa 363 
stolonifera 363 
Chrysanthemum /eucanthemum 366 
Conyza bonariensis 363f 
floribunda 363 
sumatrensis 363 
Coronopus didymus 355 
Cotula australis 363 
goughensis 363 
moseleyi 363 
Crassula pellucida 360 
Crepis capillaris 364 
Ctenitis aquilina 406 
Cynodon dactylon 386 
Cynosurus cristatus 388 
Cyperus congestus 375 
esculentus 375 
longus 376 
tenellus 376 


Dactylis caespitosa 396 
glomerata 388 
Deschampsia christophersenii 387*, 388 
mejlandii 389, 390* 
robusta 390*, 391 
wacei 392, 393* 


ambrosioides* var. 


Digitaria sanquinalis 392 
Dryopteris affinis 407 
aquilina 406 
borreri 407 
paleacea 407 
parallelogramma 407 
pseudomas 407 
tomentosa 404 
wallichiana 407 


Echinochloa crus-galli var. breviseta 392 
Edwardsia macnabiana 357 
Elaphoglossum hybridum 407 
insulare 407 
laurifolium 407 
obtusatum 407 
succisaefolium 407 
Eleusine africana 392 
indica subsp. africana 392 
Empetrum medium 367 
nigrum var. rubrum 367 
rubrum 367 
Eriosorus cheilanthoides 401 
Eucalyptus viminalis 361 
Euphorbia peplus 373 


Festuca arundinacea var. mediterranea 394 
elatior subsp. arundinacea var. genuina 
subvar. mediterranea 394 
myrus 400 
rubra subsp. commutata var. barbata 394 
Ficus carica 373 


Galicum aparine 362 

Geranium dissectum 356 

Glyceria insularis 393*, 394 
multiflora 394 

Gnaphalium candidissimum 364 
luteo-album 364 
purpureum 364 
pyramidale 364 
thouarsii 364 

Grammitis armstrongii 404 
australis var. nana 404 
billardieri 403 
billardieri var. magellanica forma nana 404 
cheilanthoides 401 
magellanica subsp. magellanica 403 
nana 404 
poeppigiana 404 

Gymnogramma cheilanthoides 401 
filipendulaefolia 401 


Histiopteris incisa var. carmichaeliana 404 
Holcus lanatus 395 
Hordeum glaucum 395 
leporinum 395 
Huperzia insularis 401 
Hydrocotyle asiatica 361 
capitata 361 


VASCULAR PLANTS FROM TRISTAN DA CUNHA 419 


Hymenophyllum aeruginosum 402 depressa forma fimbriata 362 


menziesii 403f 

peltatum 403 

peltatum var. mensiesii 403f 

tunbrigense 403 
Hypochoeris glabra 364 


Hypolepis rugosula var. villoso-viscida 404 


Isolepis bicolor 376 
squarrosa 376 
subtilis 377 
sulcata 377 
verruculosa 377 


Juncus bufonius 373 
effusus 374 
macer 374 
tenuis 374 
tristanianus 374 


Lactuca scariola 366 
serriola 366 
Lagenophora commersonii 366 
nudicaulis 366 
Leonotis leonurus 369 
Leontodon taraxacoides 366 
Lepidotis diaphana 401 
Leucanthemum vulgare 366 
Lithospermum sp. 368 
Lobelia erinus 367 
Lolium x hybridum 395 
multiflorum 395 
perenne 395 
perenne x multiflorum 395 
rigidum 395 
Lomaria alpina 409 
antarctica 409 
boryana 408 
robusta 408 
Lycopodium clavatum 401 
diaphanum 401 
insulare 401 
magellanicum 401 
saururus 401 
selago var. hessei 401 


Malus sylvestris subsp. mitis 360 
Malva parviflora 356 
sylvestris 356 
Mariscus congestus 375 
Medicago denticulata 357 
polymorpha 357 
sativa 357 
Myosotis discolor 368 


Nasturtium officinale 355 

Nephrodium aquilinum 406 

Nertera assurgens 362 
depressa 362 


granadensis 362 
holmboei 362 


Oenothera indecora subsp. bonariensis 361 


Ophioglossum opacum 401 
Oxalis corniculata 357 
purpurea 357 
variabilis 357 


Parodiochloa flabellata 396 
Paspalum dilatatum 396 
Pelargonium acugnaticum 356 
australe var. acugnaticum 356 
grossularioides 356 
Pennisetum clandestinum 398 
Peperomia berteroana 372 
tristanensis 372 
Phalaris mollis 390 
cespitosa 385 
tuberosa 398 
Phormium tenax 373 
Phylica arborea 357 
nitida 357 
Physalis peruviana 369 
Pinus canariensis 400f 
caribaea 400 
halepensis 400f 
insignis 400f 
pinaster 400f 
radiata 400f 
Plantago lanceolata 370 
major 370 
Poa annua 398 
cooki 396 
exilis 398 
flabellata 396 
infirma 398 
pratensis 398 
pratensis subsp. pratensis 398 
pratensis subsp. subcaerulea 398 
subcaerulea 398 
trivialis 399 
Polycarpon tetraphyllum 355 
Polygonun aviculare 371 
Polypodium adiantiforme 408 
aquilinum 406 


billardieri var. magellanicum 403 
billardieri var. magellanicum forma nanum 


404 
calyptratum 408 
magellanicum 403 
poeppigianum 404 
tomentosum 404 
villoso-viscidum 404 


Polypogon intermedius 399 


mollis 399 
monspeliensis 399 


420 


Polystichum adiantiforme 408 
mohrioides 406 

Ponceletia arundinacea 399 

Prunella vulgaris 370 

Prunus persica 360 

Pteris palmaeformis 408 
vespertilionis var. carmichaeliana 404 
vittarioides 402 


Ranunculus acris 354 
biternatus 354 
carolii 354 
crassipes 354 
repens 354 
Raphanus sativus 355 
Romulea rosea var. australis 373 
Rosa rubiginosa 360 
spinosissima 360 
Rostkovia magellanica 374 
tristanensis 374 
Rubus saxatilis 360 
Rumex acetosella subsp. angiocarpus 371 
angiocarpus 371 
crispus 371 
cuneifolius 371 
frutescens 371 
obtusifolius subsp. agrestis 372 
obtusifolius subsp. obtusifolius 372 
obtusifolius var. agrestis 372 
steudelii 372 
Rumohra adiantiformis 408 


Salix babylonica 373 

Scirpus bicolor var. bicolor 376 
bicolor var. virens 376 
cernuus var. subtilis 377 
chlorostachyus 377 
moseleyanus 377 
oliveri 376 
prolifer 376 
prolifero-ramosus 376 
sulcatus var. moseleyanus 377 
sulcatus var. sulcatus 377 
thouarsianus 376 
thouarsianus var. bicolor 376 
thouarsianus var. pallescens 376 
thouarsianus var. virens 376 
verruculosus 377 
virens 376 

Scleranthus annuus 355 

Senecio vulgaris 366 


E. W. GROVES 


Silene alba 355 
gallica 355 
Sisymbrium glaciale 355 
Solanum nigrum 369 
Sonchus arvensis 366f 
asper 366 
oleraceous 366 
Sophora macnabiana 357 
microphylla 357 
tetraptera forma goughensis 357 
tetraptera var. microphylla 357 
Spartina arundinacea 399 
Spergula arvensis 356 
Sporobolus africanus 400 
capensis 400 
indicus 400 
Stellaria media 356 


Tetroncium magellanicum 375 
Thelypteris tomentosa 404 
Trichomanes aeruginosum 402 
angustatum 403 
tenerum 403 
Trifolium dubium 358 
micranthum 359 
pratense 359 
repens 359 
subterraneum 359 


Uncinia brevicaulis var. brevicaulis 378 


brevicaulis var. gracilior 378 
brevicaulis var. macloviana 378 
brevicaulis var. rigida 378 
compacta var. elongata 378 
gracilis 378 
meridensis 378 
sinclairei 378 
smithii 378 

Ulex europaeus 359 

Urostachys insularis 401 


Verbascum virgatum 369 

Verbena officinalis 369 

Veronica agrestis 369 
serpyllifolia 369 

Vittaria stricta 402 
vittarioides 402 

Vulpia bromoides 400 
myuros 400 


Xeranthemum caespitosum 363 


( 


British Museum (Natural History) 
1881-1981 


Centenary Publications 


Chance, change & challenge 


Two multi-author volumes from one of the foremost scientific institutions in the world. 


General Editor: P. H. Greenwood 


The Evolving Earth 


Editor: L. R. M. Cocks 


The Evolving Biosphere 
Editor: P. L. Forey 


In the first volume, The Evolving Earth, twenty scientists have been asked to review 
the present state of knowledge in their particular field, ranging from the origin of 
the Earth, through ocean sediments and soils to continental drift and palaeogeography. 


In the companion volume, The Evolving Biosphere, museum scientists have chosen 
an evolutionary concept—speciation, coevolution, biogeography etc. and related 
this to the group of animals or plants in which they are specialising. Thus beetles 
and birds exemplify sympatric and allopatric speciation, butterflies mimicry and 
certain fishes explosive evolution. 


In both volumes the text is supplemented by over one hundred specially-commissioned 
pieces of two-colour artwork. 


These two books will be invaluable to all sixth-form and undergraduate biology and 
geology students. 


The Evolving Earth: 276219 mm, 280pp, 138 line illustrations, 42 halftones 
The Evolving Biosphere: 276x219 mm, approx. 320pp, 133 line illustrations 
Published: May 1981 


Co-published by the British Museum (Natural History), London and Cambridge 
University Press, Cambridge. 


Titles to be published in Volume 8 


The Thelypteridaceae of Ceylon. By W. A. Sledge 


Studies in the genus Hypericum L. (Guttiferae) 
2. Characters of the genus. By N. K. B. Robson 


A revision of the lichen family Thelotremataceae in Sri Lanka. 
By Mason E. Hale, Jr. 


Vascular plant collections from the Tristan da Cunha group of 
islands. By Eric W. Groves 


Printed by Henry Ling Ltd, Dorchester 


aS : 


sel eek 
a 


ee 
erste 


bgt 


tite 


te 


é 
ty 2, 


Aa 


ms 


ay .8, 


thes 


e 
:, 


Ss, 


asitatet 
tea 


¢ 


eee 


<i 
at 


eet Matas