Biological Conservation 142 (2009) 1754–1766

Contents lists available at ScienceDirect

Biological Conservation

journal homepage: www.elsevier .com/locate /b iocon

Use of the Biological Flora framework in the United Kingdom Overseas Territories:Euphorbia origanoides L.

Alan Gray a,*, Paul David Robinson b, Stedson Stroud c

a Centre for Ecology and Hydrology, Edinburgh Research Station, Bush Estate, Penicuik, Midlothian, Scotland EH26 0QB, UKb Joint Nature Conservation Committee, Monkstone House, City Road, Peterborough PE1 1JY, UKc Ascension Island Conservation, Conservation Centre, Georgetown, Ascension Island, ASCN IZZ, South Atlantic Ocean, UK

a r t i c l e i n f o a b s t r a c t

Article history:Received 5 November 2008Received in revised form 13 March 2009Accepted 15 March 2009Available online 11 April 2009

Keywords:EuphorbiaPlant conservationEndemic speciesGerminationHerbivoryReproductive biologyAscension IslandAscension spurge

0006-3207/$ - see front matter � 2009 Elsevier Ltd. Adoi:10.1016/j.biocon.2009.03.014

* Corresponding author. Tel.: +44 131 445 8471; faE-mail address: alangray@ceh.ac.uk (A. Gray).

The United Kingdom Overseas Territories (UKOTs) are globally important for a high diversity of endemicand threatened plant species but are poorly represented in plant ecological literature. This lack of ecolog-ical research is compounded by a lack of funding and skills. Cost effective approaches of compiling con-servation relevant information are required. Here we present the first examination of a species from theUKOTs presented within the standard framework of a Biological Flora. This framework allows a conve-nient way to compile ecological information and assess missing data. The account reviews all availableinformation on Euphorbia origanoides L. (Ascension spurge) from Ascension Island (South Atlantic Ocean)relevant to understanding its ecology and conservation, including soil chemistry, climate and plant com-munity data. E. origanoides is an endemic perennial, found in dry, lava plains of Ascension Island withsoils comprised of weathered volcanic scoria. E. origanoides has suffered habitat loss through the intro-duction of invasive species and survival in the wild is currently under threat. We relate the informationgathered for this Biological Flora to the conservation of the species in the wild and propose the frame-work should be used as one way of compiling information relevant for conservation managers. Theframework is beneficial as it allows an evidence-based approach to conservation but also permits the pri-oritisation of research and can help conservation managers to meet targets for the Convention on Biolog-ical Diversity and the Global Strategy for Plant Conservation.

� 2009 Elsevier Ltd. All rights reserved.

1. Introduction

The United Kingdom Overseas Territories (UKOTs) are globallyimportant for the high diversity of endemic and threatened plantspecies they contain but are poorly represented in plant ecologicalliterature. This lack of ecological research is further compoundedby a distinct lack of funding for conservation efforts. Funding deficitsare a pervasive problem in the UKOTs; at present funding mainlyconsists of the Overseas Territory Environment Programme cur-rently at about £1.5 million per year, representing a shortfall ofaround £15 million (RSPB, 2007). This deficit is augmented to someextent by programmes such as the Darwin Initiative and a number ofterritories have been successful in gaining Darwin Initiative fund-ing. As of the 31st of March 2008 the Darwin Initiative has contrib-uted over £3 million to biodiversity conservation in the PacificIslands, the Caribbean and the Atlantic Islands which include manyUKOTs (Anon, 2008). However, even this significant contribution toconservation does not address the funding shortfall. In comparison,UK public expenditure on biodiversity for 2004–2005 was estimated

ll rights reserved.

x: +44 131 445 3943.

to be around £300 million (DEFRA; http://www.defra.gov.uk/envi-ronment/statistics/supp/spkf20.htm (accessed 13.01.09)). Thefunding bias towards the UK mainland may not be the most effectiveuse of available funds, particularly given the rich biodiversity foundin the UKOTs. For example, 180 endemic plant species are found inthe UKOTS compared to 72 in mainland UK; 26 of these UK speciesare at the sub-specific level, four are doubtful and one is regarded asonly probably endemic (Stace, 1997, see also http://www.nhm.ac.uk/nature-online/life/plants-fungi/postcode-plants/checklist-british-endemic-plants.html). There are also a greater number of threa-tened species in UKOTs (nearly 10 times the number of threatenedspecies are found in the UKOTs, 2259 species, cf. 379 in mainlandUK as listed by the IUCN). This suggests that a modest provision offunding towards the UKOT’s could yield a substantial global biodi-versity gain. Undoubtedly, this would be an effective means of meet-ing some of the UKs’ obligations under the Convention on BiologicalDiversity (CBD) and towards targets of the Global Strategy for PlantConservation. In addition this would go some way to delivering theprinciples and commitments laid down in the Environmental Char-ter of each individual territory and hence territory obligations toCBD. However, the problem is not just one of finances there is alsoa lack of specialised skills and/or the specific expertise required

A. Gray et al. / Biological Conservation 142 (2009) 1754–1766 1755

for conservation programme implementation and sustainability inUKOTs.

At present funding is not forthcoming, therefore, cost effectivemethods of compiling conservation relevant information are re-quired as much to avoid wasting limited resources as to highlighttechnical capacity and applied research requirements. In 1928 Pro-fessor E.J. Salisbury put forward a proposal for a British ‘Biological’Flora; the first accounts being published in 1941 (Anon, 1941). Thespecies accounts consist of a standardised format which has re-mained relatively consistent with similar formats being adoptedworldwide (e.g. Wehi and Clarkson, 2007). Thus, it provides a well-understood framework for collating and reviewing autecologicalinformation on rare and endangered species relevant to conserva-tion. In most cases only relatively complete accounts have been pub-lished. However, for the purpose of compiling rare speciesinformation, the standard framework allows the identification ofmissing data. This can be of almost equal importance when settingconservation/research agendas on a limited budget; both in termsof finance and skills base.

Here we present the first examination of a species from theUKOTs presented within the standard framework of a BiologicalFlora. We propose that this framework is one method to allow con-servation managers not only to act on the information gathered butaid the prioritisation of research needs for threatened plant speciesand ecosystems.

2. Methods

This account reviews all available information on Euphorbiaoriganoides L. (Ascension spurge) from Ascension Island (SouthAtlantic Ocean) relevant to understanding its ecology and conser-vation. Sources have included reports, peer reviewed papers, andherbarium specimens.

3. Biological flora

3.1. Taxonomy

Family: Euphorbiaceae sensu stricto (APGII).Scientific Name: Euphorbia origanoides L. (sub-genusChamaesyce).Common Name: Ascension spurge.

E. origanoides L. is a perennial dwarf shrub. Plants form hemi-spherical domes that can extend over 1 m in diameter, Table 1.Stems are usually reddish (though sometimes only on the upperside), dichotomously branched, possessing white latex with olderwoody growth at the base. Tap roots can reach over 1 m (Gray,2003), however, this is undoubtedly an underestimate as rootsare prone to snapping; to our knowledge only one large matureplant has ever been excavated to this extent.

Stipules are broadly ovate hyaline and fringed, becoming brownwith age. Leaves are opposite, shortly petiolate, asymmetricallycordate, with distinctly toothed margins and are conspicuously3–5 nerved. Upper leaf surface is glabrous, dark green or glaucousgreen; the lower side is a paler glaucous green when fresh.

Flowers borne at the tips of branches in leaf axils, monoeciousin distinctive small units composed of one female and many maleflowers grouped together in a cup shaped cyathium with four con-spicuous glands at top. Cyathia are solitary in terminal compoundcymes with paired branches each subtended by a bract which isleafy but more elongated than the leaves. Perianth absent thoughthe four glands are creamy-yellowy-white in colour giving theappearance of true flowers. Stamens 1–3 mm long with jointed fil-ament, ovary and fruit 3 (4) celled with 1 ovule per cell. Fruit a gla-

brous schizocarp approximately 1–2 mm in diameter, seeds areapproximately 0.5–1 mm; see Fig. 1.

No known variants have been described although populationsshow some variation in height and diameter (Table 1). The extentto which this variation is environmentally or genetically determinedhas not yet been explored. Fig. 2 shows the general habit, flowers,fruits and likely pollinators including members of Hemiptera andDiptera.

E. origanoides is confined to lower more arid regions of Ascen-sion Island where soils are largely composed of weathered scoriaand volcanic debris.

3.2. Geographical and altitudinal distribution

E. origanoides is one of 10 endemic vascular species from Ascen-sion Island (97 km2) a volcanic island in the South Atlantic Ocean(7�570S, 14�220W) four of which are now considered extinct (Ash-mole and Ashmole, 2000; Cronk, 1980; Gray et al., 2005). Ascen-sion Island, discovered in 1701, was permanently settled in 1815just a few days after Napoleons’ imprisonment on St Helena (Ash-mole and Ashmole, 2000). Around 1845 a large scale programme ofplant introductions began that has resulted in large changes toAscensions’ vegetation (Ashmole and Ashmole, 2000; Cronk,1980; Duffey, 1964).

Fig. 3 shows a simplified habitat map characterised from satel-lite imagery and altitudinal vegetation zones as delimited by Stü-der (1889) which were also adopted by Duffey (1964), Gray et al.(2005) summarised the main changes since the visits of Stüderand Duffey which together with the data compiled for Fig. 3 formthe basis of the following descriptions.

Zone 1: This is an arid area of lava fields and low craters below330 m and is characteristically where E. origanoides L. is nowfound. The vegetation remains similar to the description of Duffey(1964), with very patchy but often extensive vegetation, and abun-dant annual species appearing after rainfall events especiallygrasses. In addition to E. origanoides, species present in this zoneinclude the native Aristida adscensionis L., Portulaca oleracea L.and Cyperus appendiculatus Kunth, and the introduced Tecoma stans(L.) Juss. ex Kunth, Enneapogon cenchroides (Licht.) C.E.Hubb., Arge-mone mexicanca L., Heliotropium curassavicum L., Nicotiana glaucaR.Grah., and Waltheria indica L. The most notable change since Stü-der (1889) and Duffey (1964) is the abundance of Prosopis juliflora(Sw.) DC, now dominating areas of this zone particularly in thewest (Fig. 3) appearing to have displaced much of the Acacia scrubdescribed by Duffey (1964).

Zone 2: This area lies between 330 and 660 m and has a morecomplete coverage of vegetation than zone 1 including scrub, wood-land, and grassland habitats. Characteristic species include native A.adscensionis, P. oleracea, Ipomoea pes-caprae (L.) R. Br. and Nephrol-epis hirsutula (G. Forst.) C. Presl and the introduced P. juliflora, Junipe-rus bermudiana L., Causurina equisetifolia L., T. stans, Opuntia sp.,Leucaena leucocephala (Lam.) de Wit, Lantana camara L., N. glauca,Psidium guajava L., and Acacia spp.. As described by Duffey (1964)this is the zone where the grass Melinis minutiflora P. Beauv. attainsits greatest abundance. As far as is known this area has never beenoccupied by E. origanoides (though, see Section 3.3). Mainly centredon Green Mountain, this zone and zone 3 below would have been themain areas for the extinct endemic species Anogramma ascensionis(Hook.) Diels, Dryopteris ascensionis (Hook.) Kuntze, Oldenlandiaadscensionis (DC.) Cronk, and Sporobolus durus Brongn.; and cur-rently accommodate the extant endemic species Asplenium ascen-sionis Watson Marattia purpurascens de Vriese Pteris adscensionis(Forst.) Sw., Sporobolus caespitosus Kunth, and Xiphopteris ascension-ense (Hieron.) Cronk.

Zone 3: An area often covered in mist extending from 660 to850 m and previously wrongly described as rainforest (Pearce,

Table 1Mean (SE) and maximum measurements for selected characteristics of Euphorbia origanoides plants recorded in 2003.

Measurement (mm) Site n Mean (SE) Max

Plant height Cotar Hill 203 87.4 (3.8) 330Letterbox 7 91.4 (9.1) 120Mars Bay 84 151.5 (7.1) 360Round Hill 80 82.1 (2.6) 150Spire Beach to Hummock Point 45 65.1 (4.7) 160South Gannet 1170 77.7 (1.4) 330Sisters Peak 22 26.8 (5.8) 100All data 1611 82 (1.3) 360

Plant diameter Cotar Hill 203 150.3 (0.8) 630Letterbox 7 150.1 (3.8) 330Mars Bay 84 330 (1.9) 1470Round Hill 80 120.1 (0.5) 260Spire Beach to Hummock Point 45 150.1 (1.5) 470South Gannet 1170 140.2 (0.3) 1520Sisters Peak 22 20.5 (0.5) 80All data 1611 150.1 (0.3) 1520

Branch length Mars Bay 40 72.3 (5.4) 194Sisters 7 54.2 (4.9) 82.3South Gannet 77 71.1 (3.3) 148.4All data 124 65.9 (2.7) 194

Upper leaf length Mars Bay 36 9.7 (0.4) 17.4Sisters 6 8 (0.5) 9.9South Gannet 71 8.6 (0.2) 12All data 113 8.7 (0.2) 17.4

Middle leaf length Mars Bay 38 9.7 (0.4) 13.6Sisters 6 9 (0.6) 11South Gannet 72 9.4 (0.2) 14.3All data 116 9.4 (0.2) 14.3

Lower leaf length Mars Bay 35 7.9 (0.4) 12.6Sisters 6 7.5 (0.9) 10.6South Gannet 64 8.5 (0.3) 14.5All data 105 7.9 (0.2) 14.5

Upper leaf width Mars Bay 36 6.1 (0.4) 10.8Sisters 6 5.4 (0.4) 6.3South Gannet 71 5.6 (0.2) 9.6All data 113 5.7 (0.2) 10.8

Middle leaf width Mars Bay 37 7.4 (0.2) 11Sisters 6 6.3 (0.4) 8.1South Gannet 72 6.8 (0.2) 9.7All data 115 6.8 (0.1) 9.7

Lower leaf width Mars Bay 35 6.4 (0.3) 9.5Sisters 6 6.6 (0.7) 9.6South Gannet 64 6.4 (0.2) 10.4All data 105 6.4 (0.2) 10.4

1756 A. Gray et al. / Biological Conservation 142 (2009) 1754–1766

2004). The vegetation forms an almost complete cover made up ofa mosaic of grassland, scrub, woodland, bamboo and ginger. Spe-cies within this mosaic include Alpinia zerumbet (Pers.) B.L. Burttand R.M. Smith, Bambusa sp., Podocarpus elongatus (Aiton) L’Herit.ex Pers., Auracaria excelsia (Lamb.) R. Br., J. bermudiana, L. leucocep-hala, Ficus spp., P. guajava, Sporobolus africanus (Poir.) Robyns andTournay, and Paspalum spp. This area was previously describedas a carpet of ferns (Hooker, 1867), again, E. origanoides is notknown from this zone.

Fig. 4 shows the distribution of E. origanoides populations fromsurveys in 1958, 1976, 1998 and 2003, since at least 1958 this hasbeen patchy and localised. Ashmole and Ashmole (2000) and Cronk(1980) speculate that the distribution may have been more wide-spread previous to human occupation. Over the 45 year period itis estimated that the E. origanoides population has seen a reductionof around 50% in area occupied. The current distribution is similarto that shown for 2003 but with one notable difference, the popu-lation at South Gannet Hill (presently the largest population) hasundergone an estimated 40% loss in area since 2003; this has beenoffset somewhat by resurgence in some areas e.g. along the coastalarea from Hummock Point towards Letterbox. Populations appear

to be somewhat ephemeral in nature disappearing only to reap-pear in the same or nearby sites after a period of time (Cronk,1980; Duffey, 1964) this suggests a persistent seed bank and/orperiodic seed dispersal to favourable areas. Seeds stored at roomtemperature in silica gel remain viable for at least 5 years (Grayunpublished data).

Populations at present can be found at South Gannet Hill, MarsBay, Cotar Hill, Sisters Peak, Wig Hill, Letterbox, near Powers Peakand scattered along the coastal area from Spire Beach to HummockPoint (Fig. 4). The population at Comfortless Cove noted by Grayet al. (2000) recorded dead at the time of the survey has not reap-peared, though a new plant was discovered in 2008 approximately0.5 km south of the old population. A population at the eastern endof Cross Hill recorded by Rudmose-Brown in 1906 (see Cronk,1980) has not survived to the present day. However, one deadplant was recorded by Gray et al. (2000) on the western flank ofCross Hill but again no plants have reappeared since this time.The disappearance of the population at English Bay coincided withthe construction of the BBC site in 1964 and it is likely that the dis-turbance associated with this caused the local extinction at thissite (Ashmole and Ashmole, 2000; Gray et al., 2005).

Fig. 1. Euphorbia origanoides L.: (a) fruit; (b) seeds; (c) side view of seedling showing cotyledons and first true leaves; and (d) plan view of seedling showing cotyledons andfirst true leaves.

Fig. 2. Euphorbia origanoides L.: (a) general habit; (b) flowers fruits and likely pollinators including members of Hemiptera and Diptera; (c) close up of inflorescence; and (d)grazed stem from the Hummock Point site. Scale bars are approximate.

A. Gray et al. / Biological Conservation 142 (2009) 1754–1766 1757

3.3. Habitat

3.3.1. Climatic and topographical limitationsAnnual rainfall and mean temperature data are given in Table

2 for the Airhead, Georgetown and Pan-Am Beach. Mean annual

rainfall at Georgetown between 1899 and the present is128 mm year�1, during this period the largest amount recordedwas in 1924 (471 mm year�1) and the lowest in 1970 (22 mmyear�1); rainfall events are episodic and there are months wherelittle or no rain falls hence periods of drought are not uncommon.

Fig. 3. (a) Map of Ascension showing the location of main areas of habitation, sites mentioned in the text, roads and the general location of Ascension Island in relation toAfrica and South America (inset). (b) Simplified habitat map classified from satellite imagery; delineated by the 330 m and 660 m contour lines are the vegetation zones ofStüder (1889). Habitat map reproduced with the kind permission of the Ascension Island Government, Ascension Environmental Information Operations Utility.

1758 A. Gray et al. / Biological Conservation 142 (2009) 1754–1766

The mean monthly minimum and maximum temperature datasuggest a Continentality Index of 8.8, hence E. origanoides canbe classified as a Hyperoceanic; barely hyperperoceanic species(using the criteria of Rivas-Martínez et al., 1999). The area ofAscension where E. origanoides is found can also be defined interms the UNEP index of aridity (AIu) (UNEP, 1992). From theavailable data for Ascension from the Meteorological Office atthe Airhead: AIu = 0.056. This suggests that E. origanoides inhabitsan area that almost qualifies for the upper range of Hyperarid

(AIu < 0.05) but can certainly be classified as Arid (AIu between0.05 and 0.20).

Most populations are found between sea level and 150 m, thehighest population is currently on the Sisters Peak range at330 m. Strict altitudinal limits remain unknown and E. origanoidesmay have had a wider altitudinal distribution historically. Germi-nation has been observed at slightly higher altitudes than recordedin the wild during recent re-introduction trials on Green Mountain(approximately 400 m), though survival is limited by competition

Fig. 4. Distribution of Euphorbia origanoides L. populations on Ascension Island; (a) distribution recorded by Duffey in 1958, locations are: Cross Hill, English Bay, LetterBox, MarsBay, North East Bay and Spire Beach to Hummock Point (Duffey, 1964); (b) distribution recorded by Cronk in 1976, locations are: Cotar Hill, Cross Hill, LetterBox, Mars Bay, NorthEast Bay, Pyramid Point Road, Round Hill, Sisters Peak, South Gannet, and Spire Beach to Hummock Point (Cronk, 1980); (c) distribution recorded by Gray et al in 1998 ComfortlessCove, Cotar Hill, Mars Bay, Pyramid Point Road, Round Hill, Sisters Peak, South Gannet Hill, Spire Beach to Hummock Point, and North East bay (Gray et al., 2000); and (d)distribution recorded by Gray in 2003, locations are Cotar Hill, LetterBox, Mars Bay, Round Hill, Sisters Peak, South Gannet Hill, Spire Beach to Hummock Point and Wig Hill (Gray,2003).

Table 2Selected climate statistics for the Airhead (1987–2007), Georgetown (1899–2007) and Pan-Am Beach (1977–2006) from Ascension Island, South Atlantic Ocean. Data fromAscension Island Meteorological Office.

Site Temp/rain Statistic January February March April May June July August September October November December Year mean

Airhead Temp Mean 26.2 27.2 27.8 27.8 27.1 26.1 25.2 24.5 24.2 24.3 24.6 25.4 25.9Mean Max. 28.9 29.9 30.5 30.3 29.5 28.5 27.6 26.9 26.7 27.0 27.3 28.1 28.4Mean Min. 23.5 24.4 25.1 25.3 24.7 23.7 22.9 22.0 21.6 21.7 22.0 22.7 23.3

Airhead Rain Mean 7.4 16.8 21.6 38.3 12.0 11.5 12.1 14.2 12.0 8.8 12.0 6.2 172.3Wettest 36.8 94.8 82.2 344.4 70.8 39.5 58.4 32.2 38.0 22.9 39.3 18.4 493.6Driest 0.3 0.0 0.8 3.3 0.1 0.4 1.4 0.8 1.0 1.4 3.2 1.4 67.4

Georgetown Mean 4.0 9.8 20.7 26.6 14.5 11.8 11.4 10.3 8.1 6.1 5.0 3.2 128.1Wettest 21.4 137.6 191.4 321.2 272.7 58.1 61.0 93.2 24.6 33.1 34.4 12.4 470.9Driest 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 22.2

Pan-Am Beach Mean 8.3 12.9 32.7 32.7 11.2 10.7 12.5 11.1 9.1 8.1 10.1 5.9 152.2Wettest 38.2 111.8 294.6 338.6 68.8 39.5 53.4 32.3 30.0 25.1 65.3 18.1 525.2Driest 0.1 0.0 0.0 0.8 0.0 0.8 1.0 0.3 1.0 0.0 1.8 0.1 0.0

A. Gray et al. / Biological Conservation 142 (2009) 1754–1766 1759

with introduced animal and plant taxa (Gray and Stroud, unpub-lished data). The current population distribution therefore appearsto be limited as much by competition from introduced taxa as byclimatic factors.

3.3.2. SubstratumE. origanoides is a characteristic of the low lying lava plains of

Ascension Island where soils are composed of weathered volcanicscoria. Table 3 shows soil chemistry data for the main populationssampled during 2003. Overall the data are variable, sites have med-ium to high soluble salt content, generally high phosphorus, magne-sium, sodium and pH but with correspondingly low values fororganic mater (and hence estimated nitrogen release) and calcium.The extremely high levels of soluble salts at the English Bay site per-

haps reflect the coastal nature of the site, although Comfortless Cove,also a coastal site, has comparably lower salinity levels. It may bepossible that historical deposition of guano at bird colonies in theEnglish Bay area raised salinity levels. However, in a study con-ducted on 11 midriff islands in the Gulf of California salinity wasfound not to differ between bird populated and non-bird areas (Waitet al., 2005).

There does not seem to be any preference for volcanic substrateE. origanoides being found on most of the volcanic deposits exceptthose found on Green Mountain which date from approximately 1million years ago (Nielson and Sibbett, 1996). As previously indi-cated, absence in this area may however, be more to do with in-creased competition from introduced species rather thanpreference for volcanic substrate.

Table 3Soil chemistry for eleven Euphorbia origanoides L. sites on Ascension Island, sampled in February 2003. Mean values (SD) for each site are given. Topsoil samples were taken usinga collecting tin (depth 5 cm, diameter 15 cm) and trowel. Moisture holding capacity was determined gravimetrically by weight. pH and soluble salts were determinedelectrometrically after mixing a 1:2 dilution of soil to distilled water set for 1 h. Organic matter was extracted with sodium dichromate and determined colorimetrically using aGilford Spectrophotometer. Estimated nitrogen release is determined from empirically derived relationship where: estimated nitrogen release = 43.34 + 20.12 *organic matter.Phosphorus results were extracted using P1 and P2 Bray solutions and Olsen’s method and determined colorimetrically using a Gilford Spectrophotometer. Potassium,magnesium, calcium and sodium were extracted using neutral ammonium acetate and determined by atomic absorption. Cation exchange capacity was calculated fromdetermined cation values.

Soil nutrientmean (SD)

Site (n)

ComfortlessCove (5)

CotarHill (10)

CrossHill (5)

EnglishBay (5)

Letterbox(10)

Mars Bay(10)

PyramidPoint (5)

RoundHill (10)

SistersPeak (10)

SouthGannet Hill(20)

Spire Beach toHummock Point (10)

All data(100)

Moisture holdingcapacity (%)

35.01 (5.00) 35.14(6.99)

39.46(5.59)

30.06(4.58)

31.57(6.77)

52.01(19.94)

34.66(3.40)

33.60(3.50)

24.35(2.87)

33.04 (5.81) 27.27 (4.04) 33.96(10.40)

Organic matter(%)

1.34 (0.18) 2.86(2.96)

1.16(0.81)

0.64 (0.18) 0.51(0.26)

3.75(4.14)

1.18(0.08)

1.03(0.42)

0.65(0.18)

2.93 (2.48) 0.53 (0.22) 1.73(2.23)

Est. N release(kg ha�1)

79.30 (4.07) 113.34(66.39)

74.59(18.87)

62.72(4.95)

109.54(158.44)

133.06(92.86)

73.47(5.11)

72.02(9.92)

63.17(5.01)

114.74(55.68)

59.81 (6.09) 92.55(68.98)

P1 Bray(mg kg�1)

45.14 (7.01) 12.10(13.41)

6.70(5.17)

170.49(18.63)

47.41(10.62)

12.54(10.09)

37.30(4.66)

10.18(7.52)

5.76(3.48)

14.52 (9.95) 37.18 (20.04) 25.14(22.88)

P2 Bray(mg kg�1)

54.14 (2.41) 38.69(21.45)

21.73(10.04)

237.49(9.62)

62.33(13.30)

26.79(10.51)

45.64(1.65)

44.57(21.65)

18.12(7.99)

42.52(13.13)

56.79 (21.31) 46.18(24.02)

P Olsen(mg kg�1)

5.01 (0.94) 20.48(11.45)

6.21(2.80)

107.93(53.90)

17.12(11.54)

15.23(11.68)

5.55(3.08)

15.41(11.81)

4.06(0.50)

12.83 (6.47) 9.07 (7.07) 14.30(13.12)

K (mg kg�1) 11.88(10.30)

263.89(163.46)

258.59(13.19)

1055.81(468.61)

94.22(30.91)

194.79(80.10)

25.97(15.82)

368.71(58.46)

78.01(30.71)

493.40(253.19)

67.92 (49.44) 231.02(170.96)

Mg (mg kg�1) 8.35 (3.76) 151.18(91.21)

88.02(17.39)

1865.26(533.16)

19.34(6.24)

147.33(48.27)

8.44(3.05)

151.77(44.76)

21.47(3.35)

214.02(55.94)

30.77 (15.21) 131.53(156.69)

Ca 2+ (mg kg�1) 45.18(21.03)

552.49(202.54)

276.91(17.55)

1673.65(303.33)

135.17(221.87)

306.21(68.18)

37.00(10.70)

394.22(98.50)

56.90(6.92)

531.90(91.10)

78.53 (37.47) 301.45(220.04)

Na (mg kg�1) 40.81(18.44)

567.78(326.16)

218.68(143.05)

3747.47(2486.21)

81.14(37.76)

487.05(107.39)

37.34(9.86)

407.09(203.99)

90.04(45.45)

540.66(175.02)

96.03 (139.89) 364.74(395.07)

Soluble salts(mm hoscm�1)

0.59 (0.22) 3.10(1.41)

2.57(1.06)

25.34(9.35)

1.06(0.41)

4.54(2.71)

0.78(0.13)

2.37(1.18)

1.09(0.38)

1.92 (0.52) 1.34 (1.33) 3.20(5.66)

pH 6.36 (0.47) 8.04(0.90)

7.60(0.83)

5.66 (0.60) 6.50(0.63)

6.73(1.17)

7.06(0.17)

7.91(0.32)

7.49(0.31)

7.28 (0.94) 6.96 (0.66) 7.15(0.94)

CEC (meq100 g�1)

2.92 (0.90) 28.45(6.71)

27.48(5.14)

70.32(21.18)

8.49(7.98)

36.68(6.04)

2.88(0.31)

27.77(3.90)

6.41(1.67)

22.43 (5.64) 4.63 (4.26) 20.91(17.31)

1760 A. Gray et al. / Biological Conservation 142 (2009) 1754–1766

3.4. Communities

E. origanoides is an endemic component of the arid lava flowcommunities. This species poor xeric vegetation typically compriseof widely spaced plants where competition for water resources andnutrients are likely to be key (cf. Noy-Meir, 1985). Fig. 5 illustratesa Detrended Correspondence Analysis (DCA) of vegetation samplestargeted towards endemic and native vascular species (Gray et al.,2000). Fig. 5 suggests that the vegetation containing E. origanoidesis quite unlike the other vegetation sampled on Ascension. The gra-

Fig. 5. Axes 1 and 2 of a Detrended Correspondence Analysis (DCA) of percentagecover data from 143, 1 m2 vegetation quadrats. Data were recorded between Julyand August 1998 in areas where selected endemic and native vascular plants werefound (Gray et al., 2000). The endemic species are Euphorbia origanoides L. Pterisadscensionis Swartz, Sporobolus caespitosus Kunth, Asplenium ascensionis S. Watson,Marattia purpurascens De Vriese and Xiphopteris ascensionense (Hieronymus) Cronk;native species are Ipomoea pes-caprae Roth. Nephrolepis hirsutula (G. Forst.) C. Presland Histiopteris incisa (Thunb.) J. Sm. The ordination diagram shows speciescentroids (triangles) projected into the sample ordination.

dient length of axis 1 is long, 11.1 (axis 2, 5.1) with E. origanoidessamples closely clustered towards the left of Fig. 5. Although thisis partly a reflection of sampling bias towards endemic/native spe-cies distributions, the correlation with altitude also indicates sep-aration on ecological grounds with important biotic and abioticfactors such as soil, climate, and herbivore interactions accountingfor the gradient. Gradient lengths for E. origanoides samples aloneare very short (axis 1, 1.1 and axis 2, 0.7) suggesting that there isa very limited suite of species adapted to the conditions withinwhich E. origanoides is found. Mean species richness per sample,doubles from E. origanoides samples (mean = 4, SE 0.4) to those ofthe higher mountain area e.g. S. caespitosus samples (mean = 8;SE, 0.5). This is indicative of an increase in habitat richness andmay also indicate that niche complexity increases with altitudedue to e.g. a diversity of light conditions, increased moisture, theavailability of nutrients and soil formation processes. The vegeta-tion composition of Ascension Island is still in a state of flux andis as much characterised by the presence of introduced species asindigenous species (Gray et al., 2005). Nevertheless, Figs. 3 and 5indicate that Stüder’s (1889) delimitation of vegetation zones byaltitude generally appears to hold (see Section 3.2 above).

3.5. Response to biotic factors

3.5.1. GrazingLatex production is common to members of the Euphorbiaceae

and is sometimes poisonous (Heywood, 1998). When damaged, E.origanoides produces copious amounts of whitish latex from bothstems and leaves. However, as is shown in Fig. 2, latex productiondoes not appear to constitute an effective grazing deterrent. Histor-

Fig. 6. Windrose diagram of wind speed and direction from Ascension Island in theSouth Atlantic Ocean, data from Ascension Island Meteorological Office 1988–2007.

A. Gray et al. / Biological Conservation 142 (2009) 1754–1766 1761

ically, E. origanoides may have been browsed by goats, andalthough goats were eradicated in 1944, Ashmole and Ashmole(2000) hypothesize that this may partly account for the disjointednature of present populations (Fig. 4). In the lower xeric zoneinhabited by E. origanoides, annual grasses respond quickly to spo-radic rainfall events, food availability is therefore also partly peri-odic. A perennial species such as E. origanoides represents apotential food item in times of food scarcity, and although thenutritional quality is unknown, herbivores do consume stems,leaves and flowers; sheep, rabbits and mice have all been recordedeating E. origanoides. There is no experimental evidence on therecovery of E. origanoides after grazing but field observation sug-gests that unless the entire plant, is repeatedly defoliated recoverycan occur.

Information on seed predation and the identity of taxa likely topredate E. origanoides seeds is currently lacking.

3.5.2. Other plantsAs stated above, the plant communities on Ascension are as

much characterised by the presence of introduced species as indig-enous species (Gray et al., 2005). The original complement of na-tive species is thought to number around 25 vascular species.This contrasts with the present situation of nearly 350 species ofvascular plant many of which were introduced in the 19th century,although introductions continue to the present day (Ashmole andAshmole, 2000; Cronk, 1980, 2000; Duffey, 1964; Hooker, 1867).These include many species regarded as the worst invasive speciesin the world (IUCN, 2008). Most of these species are in zones 2 and3, however, Gray et al. (2005) record the presence of at least sixspecies which they consider capable of dominating vegetation inE. origanoides areas and hence threatening the survival of this spe-cies in the wild, these are: Heliotropium curassavicum L., Nicotianaglauca Graham, Leucaena leucocephala (Lam.) de Wit, Melinis minu-tiflora Beauv., Prosopis juliflora (Sw.) DC and Psidium guajava L.. Theexpansion of P. juliflora in much of the lower lava plains on Ascen-sion since the 1980’s is testament to the rapid expansion and dom-ination that some species can attain (Ashmole and Ashmole, 2000;Gray et al., 2005). A recent re-introduction trial in vegetation zone2 suggests that the quick attainment of an Ageratum conyzoides L.canopy may limit the potential for germinated seeds of E. origano-ides to survive. Fully grown planted specimens also appear to dopoorly perhaps due to similar competitive effects. This is furtherexacerbated by the frequent presence in this zone of land-crabsthat clip branches, though, crabs have not yet been observedingesting the plant material (Gray and Stroud unpublished data).This suggests that there may be competitive limitations to theareas in which E. origanoides can survive. Whether this is due tocompetition for light, nutrients or due to allelopathic effects isuncertain. However, as vegetation appears somewhat zonal, alti-tude may be one indicator for delimiting areas for future popula-tion restoration activities.

3.6. Responses to the environment

3.6.1. GregariousnessWhere E. origanoides is found it can be the dominant or co-dom-

inant component of the vegetation, on a range of volcanic soils andsand; plants readily germinate and survive on a range of composttypes in cultivation suggesting a non-specificity to soil type.

3.6.2. Performance in various habitatsE. origanoides occurs at present in only one habitat type (Fig. 5)

in the lowest vegetation zone (Fig. 3). Nevertheless, Table 1 indi-cates some difference in plant height and diameter at extant sites(see Section 3.7 below). There is at present however, no evidence todetermine whether this variation is genetic or phenotypic; there

are undoubted site differences in environmental conditions. Theconstant strong south-east trade winds (Fig. 6) may suggest a po-tential barrier for pollinators thus indicating a potential barrierto gene flow; there is also a strong environmental gradient corre-lated with altitude. These hypotheses could be tested with futuremolecular research.

3.6.3. Effect of frost drought etc.Frost does not occur on Ascension Island (Table 2), though, peri-

ods of drought are common. Plants develop long tap roots and maytherefore be able to access groundwater sources during drought.However, in other desert areas nutrient limitation during droughtappears to affect survival (Noy-Meir, 1985). It is unknown whetherthis is a contributory factor to the recent losses of E. origanoides atSouth Gannet Hill but there may be an interaction with increasedgrazing pressure as well as decreased recruitment due to less effec-tive rainfall (taking account of evapo–transpiration) in recentyears.

3.7. Structure and physiology

3.7.1. MorphologyE. origanoides typically forms a hemispherical dome (Fig. 2) with

older stems becoming woody. Early growth stems are few and up-right and continual stem breakage (through for example wind ac-tion) and abortion of the apical bud are likely to lead to this growthform.

Data on plant height, plant diameter and branch length appearto show differences between populations, for example, Sisters Peakconsistently has the lowest values (Table 1). Leaf length and widthappear to be more consistent between the three measured popula-tions. The Sisters Peak population is found at the highest altitude,approximately 330 m, and it may be that apparent differencesare due to environmental variation, however as stated above, ge-netic differences between populations cannot yet be ruled out.Examination of leaf material from unspecified populations indi-cates a mean leaf area of 126.7 mm2 (SD, 35.0; n, 386), mean dryweight of 9.9 mg (SD, 0.08; n, 382), a mean leaf dry matter contentof 29% and a mean specific leaf area of 12.9 mm2 mg�1 (Gray,unpublished data). Inspection of 10 leaf impressions from 10 sep-arate glasshouse plants showed leaves to be amphistomatic. How-ever, they appear to have a greater density of stomata on the uppersurface (193 mm�1; SD, 26) rather than the lower surface

1762 A. Gray et al. / Biological Conservation 142 (2009) 1754–1766

(86 mm�1; SD, 12). Conversely mean stomatal aperture lengthswere shorter on the upper leaf surface (0.007 mm; SD, 0.0013)compared to the lower leaf surface (0.010 mm; SD, 0.0016). Mostamphistomatous species have a greater density of stomata on thelower surface (Beerling and Kelly, 1996; Mott et al., 1982; Salis-bury, 1928) though the difference in stomatal aperture lengthmay mean that stomatal conductance is similar for both surfaces.The greater density of stomata on the upper surface may be partlybecause E. origanoides holds its leaves at an oblique angle to theincident sunlight (Cronk, 1980) this stomatal distribution may,therefore, help to avoid excessive water loss.

The fine lateral root system of E. origanoides spreads horizon-tally from the main tap root and is distributed throughout thelength of the tap root (Gray, 2003). However, the extent of lateralroot spread is difficult to ascertain since root excavation from thevolcanic scoria derived soils results in much of the root biomassbeing lost due to abrasion. This is one of the primary reasonswhy plant translocation cannot be used as a conservation measure;the destruction of the lateral root system during excavation leadsto plant death (Gray, 2003). Gray (2003) reports a mean root toshoot biomass ratio of 1:3 (n, 24) but in mature individuals the ra-tio can reach 1:10 (Gray, 2003). Gray (2003) attempted to use thesedata to show a shift in resource allocation with age towards greaterallocation to shoot biomass at around the time of first flowering.Gray (2003) speculates that this may coincide with locating avail-able ground water by plant roots. However, caution is requiredwhen interpreting these data as the sample size is small and theestimates of root biomass are undoubtedly underestimates dueto the aforementioned problem of root loss in the soil.

3.7.2. MycorrhizaNo data are available for E. origanoides (or any other species on

Ascension) though, members of the Euphorbia genus are known tohave Vesicular–Arbuscular Mycorrhizal (VAM) association. VAMhave been recorded on the roots of the closely related E. trinerviaSchumach. and Thonn. (under the synonym E. glaucophylla Poir)(Thoen, 1987), and the introduced E. hirta L. (Wang and Qiu,2006). In general VAM association in Euphorbiaceae appears tobe quite high (Harley and Harley, 1987; Trappe, 1987; Wang andQiu, 2006); in addition, significant members of the E. origanoideshabitat namely Waltheria indica L., Aristida adscensionis L., Leucaenaleucocephala (Lam.) de Wit and Prosopis juliflora (Sw.) DC., have alsobeen shown to possess mycorrhizal associates (Koske et al., 1992;Muthukumar and Udaiyan, 2000; Varma and Hock, 1999; Wangand Qiu, 2006). Further research is required to clarify if E. origano-ides and these species host mycorrhiza on Ascension Island.

3.7.3. Perennation: reproductionPlants are long lived (years) but more detailed data on longevity

are unavailable. E. origanoides flowers almost continuously untilthe onset of senescence. There are no above ground hibernatingbuds therefore E. origanoides cannot be considered a true Chamae-phyte and should perhaps be considered a perennial Therophyte(Raunkiaer, 1934). It is unknown whether plants can regeneratefrom rootstock; observations of plants growing up through oldplants may be suggestive of this but it is also likely that this is ger-mination from the seed bank in the relatively sheltered environ-ment of the dead plant.

3.7.4. ChromosomesMaterial from Wig Hill was found to be 2n = 12.

3.7.5. Physiological data3.7.5.1. Response to shade. E. origanoides inhabits the open dry lavaplains and has never been found in shade in the wild. The abovepoor germination and survival performance under an A. conyzoides

canopy may suggest intolerance to dense shade; however, this re-mains speculative since no experimental data are available.

3.7.5.2. Water relations. The only data available suggest that germi-nation can be stimulated by increased water supply (Gray, 2003).Four 1 m2 plots were given water in addition to atmospheric rain-fall for a period of 10 weeks, germination events were recordedevery week for the first 10 weeks and then again after 6 months.When compared to eight un-watered plots there was a marked in-crease in germination events; 53 in watered plots compared to 3 inun-watered.

3.7.5.3. Response to nutrients. No data are available.

3.7.6. Biochemical dataNo biochemical data are available.

3.8. Phenology

Ascension Island displays little seasonality; temperature (Table2) and light remain relatively constant year round and there is nopronounced rainy season. Flowering and fruiting appear to con-tinue throughout the life of a plant only waning at senescence.Environmental cues for particular biological processes are there-fore more difficult to ascertain and may be more closely linkedto the availability of water and nutrients for growth. Germinationmay be partly triggered by periodic large rainfall events (seeabove).

3.9. Floral and seed characters

3.9.1. Floral biologyFlowers are monoecious, and typical of the Euphorbiaceae com-

posed of one female and many male flowers together forming cya-thia resembling true flowers. Typical pollinators for Euphorbiaceaeinclude Diptera and members of this family are common among E.origanoides flowers (Fig. 2). Other likely pollinators include mem-bers of the Hemiptera which can be found in large numbersamongst plants at all sites. Controlled pollinations at Edinburghindicate that E. origanoides is self-compatible.

3.9.2. HybridsNone known.

3.9.3. Seed production and dispersal3.9.3.1. Seed production. Normally three seeds are produced perfruit though fruits with four have been observed. Fruit productionis copious and occurs throughout most of the life of the plant. Thenumber of fruits found sheltered beneath plants can also be relatedto plant size (Fig. 7). The Millennium Seed Bank reports a mean1000 seed weight for E. origanoides as 1.78 g. However, the fourvalues reported are very variable; 0.21 g 0.66 g, 1.56 g and 4.67 g(Kew, 2008). The large value of 4.67 g seems extreme, however,it is uncertain from the available data whether reported values in-clude for example, minor covering structures, debris, the surround-ing fruits or whether fresh or dry seeds were weighed (Kew, 2008).A mean value of 0.81 g excluding this extreme value would seem amore cautious estimate.

3.9.3.2. Seed dispersal. It is not known whether animal taxa areresponsible for dispersal, though the lack of an elaiosome suggeststhat ants are unlikely to disperse seeds. Other main dispersalagents are likely to be wind and water. Ripe fruits are easily trans-ported by the wind and the hemispherical dome life-form appearsto offer shelter for detached fruits and seeds; the number of fruitsfound in soil drops extremely sharply with distance from plant

Fig. 7. Number of fruits locate beneath plants in relation to plant diameter,R2 = 0.518, p < 0.001, and n = 79.

A. Gray et al. / Biological Conservation 142 (2009) 1754–1766 1763

(<1 m) (Gray, 2003). Plants have been observed growing in oldwater courses (fruits readily float). This may be partly a responseto increased available water for germination but it may also be thatseeds underneath plants are washed out during periodic heavyrains to germinate later.

Ascension is one of the remotest islands in the South Atlantic,E. origanoides is regarded as a neo-endemic or type three relict spe-cies, that is a species that has undergone recent evolutionarychange since colonisation (Cronk, 1992). The question remains thenof how the ancestor of E. origanoides reached Ascension. E. trinervia,regarded as the closest relative of E. origanoides, occurs on the coastof West Africa. Dispersal from this region to Ascension must eitherhave been by sea or air, perhaps attached to birds. Bird candidatesmay include sooty terns (Onychoprion fuscatus L.) and the AscensionIsland frigate (Fregata aquila L.) which are known to range off thewest coast of Africa. However, the fruits and seeds lack any specia-lised means for attachment to animals, though contact with the la-tex could potentially help to stick fruits and seeds to animals. It isunknown how long seeds remain viable in the ocean environment;but the prevalence of both E. origanoides and E. trinervia in coastalhabitats, and the high salinity of the soils where E. origanoides isfound, suggest a degree of adaptation to high salinity.

3.9.4. Viability of seeds: germinationThe Millennium Seed bank holds a collection of E. origanoides

seed collected from above ground live material; germination fromthis collection under standard laboratory conditions was estimatedat approximately 33% (Alton, pers. comm.). However, germinationin the field undoubtedly happens under a range of conditions and itis likely that the viability of seeds from the seed bank is the moreimportant statistic in terms of population ecology and conserva-tion management. Seeds will readily germinate under cultivationand in a range of compost and soil materials observed both inthe Ascension Island nursery and under glass in Edinburgh in theUK. Seeds collected from the seed bank have germinated after5 years storage at room temperature in silica gel (Gray, 2009). To-gether with the ephemeral nature of populations this suggests thata persistent seed bank may be a dispersal mechanism throughtime. This may be an advantageous adaptation in a habitat wheredrought is common; longer term (>5 years) viability of the seed re-mains questionable at present.

Percentage viability for seeds collected from the seed bank(stored in silica gel at room temperature) and germinated in seedcompost ranges from 0% to 2% with a mean of 0.49% for materialcollected in 2008 and 0.10% for material collected in 2003 (Gray,2009). The reduction in seed bank viability in comparison to theabove ground live material is likely to be due to processes suchas predation, desiccation, disease, and extremes of temperatureand moisture all of which require quantification.

3.10. Herbivory and disease

3.10.1. Animal feeders or parasitesAs stated above sheep, rabbits and mice have been observed con-

suming E. origanoides. Sheep graze stem, leaf and flowers; rabbitdroppings from the South Gannet area contain E. origanoides leafmaterial and the gut contents of the house mouse (Mus musculusL.) have included flowers (personal observations). Historically goatsmay have browsed E. origanoides leading Ashmole and Ashmole(2000) to postulate that this may account for the patchy modern dis-tribution. Donkeys are also present on Ascension but have not beenobserved browsing E. origanoides. Endemic species seem to be par-ticularly vulnerable to herbivory and may lack particular defencemechanisms (Bowen and Van Vuren, 1997) presumably due to a lackof selection pressure. E. origanoides, latex does not seem to be aneffective deterrent against current herbivores. There may, however,be a differential species response to compounds in Euphorbiaceaelatex. For example, consumption of the latex producing E. esula L.only appears to affect certain species, e.g. cattle; sheep and goatsare unaffected (Lym, 1998). It may be particularly unfortunate thatthe introduced herbivores on Ascension Island may not be suscepti-ble to any toxicity that E. origanoides latex possesses.

Large numbers of invertebrates are often found within the can-opy of E. origanoides, although the level of diversity of this elementof the community has not been quantified, members of Hemipteraare particularly evident (Fig. 2). The introduced cottony-cushionscale insect Icerya purchasi Mask. has been recorded on E. origano-ides and was thought to have caused considerable damage to theCross Hill and Sisters Peak populations in 1976 (Cronk, 1980). I.purchasi does not seem to have caused the widespread destructionto populations feared by Cronk (1980) but the beetle Rodolia cardi-nalis may have kept this species in check. R. cardinalis was intro-duced to St Helena in 1896 and 1898 as a biological control for I.purchasi and it is believed to have been introduced to Ascensionfor the same purpose, though the date of introduction to Ascensionis unclear (Ashmole and Ashmole, 2000).

3.10.2. Plant parasitesNone known.

3.10.3. Plant diseasesNone known.

3.11. History

Although Peter Mundy made some reference to grasses andrushes from his visit in 1656 he made no mention of E. origanoides(see Ashmole and Ashmole, 2000). The first account appears tohave been that of James Cunnihamme who recorded E. origanoides(and five other species) under the name Chamaesyce frutescenshumilior floribus comosis (Cuninghame, 1699). Linnaeus first pub-lished and described E. origanoides (though no particular holotypewas designated) in Species Plantarum (Linnaeus, 1753), from mate-rial collected by Osbeck on his journey to China, the Linnaeusdescription is as follows:

Dichotome (umbella bifida aut nulla). Euphorbia dichotoma,foliis ferrulatis ovatis obtulis trinerviis, panicula terminali, caul-ibus simplicibus. Habitat in insula Ascensionis. Planta reset itaOriganum, ut primo intuitu nequeat non pro co agnosci.

Most visitors to the Island have made some mention ofE. origanoides (see Duffey, 1964), with one notable exception,Charles Darwin. Darwin appeared to be uninterested by Ascen-sion’s flora or fauna whilst visiting during his Beagle voyage, hisattention being diverted by endeavours to describe Ascension’sgeology (Darwin, 1839).

1764 A. Gray et al. / Biological Conservation 142 (2009) 1754–1766

3.11.1. PhylogeographyNo molecular data are available for E. origanoides therefore we

can only speculate on phylogeographic relations. However, infer-ence can be drawn from a number of molecular phylogenies nowavailable for other island endemics, for example, from Hawaii,(Baldwin et al., 1991; Kim et al., 1998; Lawton-Rauh et al., 2003,2007; Wagner and Funk, 1995) the Juan Fernandez Islands, (Craw-ford et al., 2001; Sang et al., 1994) the Galapagos (Schilling and Pa-ner, 2002), Macaronesia (Böhle et al., 1996; Francisco-Ortega et al.,1996, 2000; Juan et al., 2000; Silvertown, 2004) and Ascensions’closest neighbour St Helena (Eastwood et al., 2004). These datasuggest that in most cases species are monophyletic and likely tohave diverged from a single colonisation event. The evidence fromSt Helena not only suggests that the four species of Commidendrum,C. spurium, C. robustum, C. rotundifolium, and C. rugosum, form a clo-sely related monophyletic group but also together with Melanoden-dron integrifolium may have evolved from a common ancestor thatarrived in St Helena via a single dispersal event (Eastwood et al.,2004). The closest relatives of Commidendrum and Melanodendronappear to be South African (Eastwood et al., 2004). Given prevailingwind and sea directions and that the putative relative of E. origano-ides is also African; this may suggest a similar dispersal pathwayfor species arriving on both St Helena and Ascension Island.Although, the colonisation event that gave rise to E. origanoides islikely to be a much more recent event than those on St Helena. Evi-dence from other island studies also suggests that in the main, is-land species appear to have lower genetic diversity, particularlyendemic species, than more widespread and continental species(Crawford et al., 2001; Francisco-Ortega et al., 2000; Frankham,1997). This may also be the case for E. origanoides, however, the de-gree to which E. origanoides has gone through population expan-sion, bottlenecks, and migration remains unknown.

Silvertown (2004) suggests that after the two main stages ofcolonisation (dispersal and establishment) there are two hypothe-ses to explain the large spread monophyly in island species either:dispersal barriers make repeated colonisation unlikely or, the firstsuccessful coloniser limits the establishment of late arriving colo-nists through inter-specific competition. Although Silvertown(2004) offers the latter as an explanation for the flora of the Canar-ies, we suggest that the first hypothesis is a more likely explana-tion for the endemic flora of Ascension including E. origanoides.Although repeated colonisations cannot be ruled out, andassuming no extinction before human settlement, evidence to sup-port this includes the low species diversity of Ascension, its geolog-ical youth and apparent niche availability. The depauperate natureof the vascular flora of Ascension suggests there have only beenapproximately 25 successful colonisation events in the 1 millionyear history of Ascension (very roughly 1 species every40,000 years); the majority of these are also spore-dispersingspecies (Cronk, 1980). The dispersal distances to Ascension Islandare long, thus an effective dispersal barrier is in place making col-onisation an unlikely event. The similarly remote St Helena has amore diverse flora than Ascension but St Helena is a much older is-land (Cronk, 2000). The success of introduced species during themore recent history of Ascension may be testament to availableniche space and/or reduced competition levels pre-human settle-ment. Testing these hypotheses for E. origanoides awaits moleculardata.

4. Discussion and conclusion

In terms of conservation status, Gray et al. (2005) categorised E.origanoides as Critically Endangered under current IUCN criteria,due to the presence of invasive species, the localised distributionand a fluctuating population size. These factors remain but active

measures are now in place to address the problem with an islandnursery and an ex-situ programme in the UK. The aim of this pro-gramme is to attain a viable population of E. origanoides in the wildby careful re-introduction trials taking into account climate changeand the effects of invasive species. The above information gatheredin the form of this Biological Flora is an extremely useful startingpoint for this programme. The Biological Flora framework in thiscontext is as much about missing information; for E. origanoidesthe following have no or very minimal data.

1. Mycorrhiza: Inoculation of mycorrhiza may help to establishplants in experimental trials and should be an important con-sideration in the conservation of endangered species (Sharmaet al., 2008).

2. Response to nutrients: In combination with information onmycorrhiza, response to nutrients may be crucial in determin-ing not only plant responses but also in the identification ofareas for restoring populations where nutrient deficiency maybe a problem. Competitive effects could also be tested by theuse of an experimental approach to this question.

3. Biochemical data: Although this may not help E. origanoides bio-chemical data may be informative in deciding the propensityfor herbivory. In some species these data may also be usefulin an ethno-botanical sense.

4. Hybrids: E. origanoides hybridisation is not known at present butthere are several related species on Ascension and introductionsare still occurring (Gray et al., 2005) therefore it is possible thata future introduced relative may ‘dilute’ genetic integrity.

5. Plant parasites and diseases: Although none are known at pres-ent this due to absence of data rather than the absence of par-asites or diseases. There is also the potential for introducedparasites and diseases to become apparent.

In terms of the restoration programme we suggest that the pri-ority for the next phase of research on E. origanoides should be theinitiation of experimental trials to investigate responses to nutri-ents, competitive effects and mycorrhizal relationships.

Although we have framed this account for one particular spe-cies, many of the issues raised are common to other species inthe UKOTs, such as, a focus on the control introduced species andcarefully targeted monitoring of populations. The framework mayseem to be species centric which some consider somewhat contro-versial (Bifolchi and Lodé, 2005; Delibes-Mateos et al., 2007; Hesset al., 2006; Simberloff, 1998; Smith and Zollner, 2005) but tar-geted ‘charismatic’ species are used as a focus for publicity (Farrieret al., 2007). In addition, even when plant conservation adopts anecosystem approach, in many cases endemic, rare and threatenedspecies are likely to form the focus of programmes to monitor con-servation outcomes. It is therefore vital that plausible reasons forspecies declines or increases can be formulated; information inthe format of a Biological Flora can aid this interpretation.

We therefore propose that regardless of the particular conserva-tion approach the Biological Flora framework provides a flexibleand informative way to represent plant ecological informationfor effective conservation management. We also suggest that con-servation managers in the UKOTs, where information is lacking,adopt this framework for guiding the baseline information gather-ing process of plant species conservation programmes.

The UK government are likely to fail to meet current obligationsto the CBD by 2010 (Yeo et al., 2008). Given the modest resourcesrequired and the rich biodiversity in UKOTs we suggest that theinvestment in UKOTs research, conservation and training wouldbe an effective means of meeting CBD obligations, targets of theGlobal Strategy for Plant Conservation and the UKOTs Environmen-tal Charters. One existing mechanism for this to be realised wouldbe the Darwin Initiative which is already equipped to help meet

A. Gray et al. / Biological Conservation 142 (2009) 1754–1766 1765

not only commitments to CBD but also to address education andtraining shortfalls within host countries.

Acknowledgements

We would like to thank Dale Hawkins, Gary Milne, Frances Dix-on, Lucinda Kirk, Zoe Smolka, Lucy Webster, Sam Gardner, SusannaMusick, and Tara Pelembe for help on and of Ascension. Ian Blackand the staff of Ascension Island Meteorological Office kindly pro-vided access to weather data. Julia Wilson and Stephen Cavershelped with useful discussion and comments on the text. Fundingfor field work was provided the United States Air Force, ComputerSciences Raytheon, Ascension Island Government and by the DavisFund of the University of Edinburgh. We would like to thank AlanMills for setting up the Ascension Environmental InformationOperations Utility (AEIOU) GIS database which helped create thehabitat map, the data were used with the kind permission of theAscension Island Government. We would also like to acknowledgethe anonymous referees for comments that measurably improvedthe manuscript.

References

Anon, 1941. Foreword. The Journal of Ecology 29, 356–357.Anon, 2008. Protecting Biological Diversity Around the World. The Darwin Initiative

Eleventh Annual Report.Ashmole, N.P., Ashmole, M., 2000. St Helena and Ascension Island: A Natural

History. Anthony Nelson, Oswestry.Baldwin, B.G., Kyhos, D.W., Dvorak, J., Carr, G.D., 1991. Chloroplast DNA evidence for

a North American origin of the Hawaiian silversword alliance (Asteraceae).Proceedings of the National Academy of Sciences of the United States ofAmerica 88, 1840–1843.

Beerling, D.J., Kelly, C.K., 1996. Evolutionary comparative analyses of therelationship between leaf structure and function. New Phytologist 134, 35–51.

Bifolchi, A., Lodé, T., 2005. Efficiency of conservation shortcuts: an investigationwith otters as umbrella species. Biological Conservation 126, 523–527.

Böhle, U.R., Hilger, H.H., Martin, W.F., 1996. Island colonization and evolution of theinsular woody habit in Echium L (Boraginaceae). Proceedings of the NationalAcademy of Sciences of the United States of America 93, 11740–11745.

Bowen, L., Van Vuren, D., 1997. Insular endemic plants lack defenses againstherbivores. Conservation Biology 11, 1249–1254.

Crawford, D.J., Ruiz, E., Stuessy, T.F., Tepe, E., Aqeveque, P., Gonzalez, F., Jensen, R.J.,Anderson, G.J., Bernardello, G., Baeza, C.M., Swenson, U., Silva O, M., 2001.Allozyme diversity in endemic flowering plant species of the Juan FernandezArchipelago, Chile: ecological and historical factors with implications forconservation. American Journal of Botany 88, 2195–2203.

Cronk, Q.C.B., 1980. Extinction and survival in the endemic flora of Ascension Island.Biological Conservation 17, 207–219.

Cronk, Q.C.B., 1992. Relict floras of Atlantic Islands: patterns assessed. BiologicalJournal of the Linnean Society 46, 91–103.

Cronk, Q.C.B., 2000. The Endemic Flora of St Helena. Anthony Nelson, Oswestry.Cuninghame, J., 1699. A catalogue of shells, etc. Gathered at the Island of Ascension,

by Mr. James Cuninghame. Philosophical Transactions (1683–1775) 21, 295–300.

Darwin, C., 1839. Journal of Researches into the Geology and Natural History of theVarious Countries Visited by H.M.S Beagle Under the Command of CaptainFitzroy R.N. from 1832 to 1836, New ed. J. Murray, London.

Delibes-Mateos, M., Redpath, S.M., Angulo, E., Ferreras, P., Villafuerte, R., 2007.Rabbits as a keystone species in southern Europe. Biological Conservation 137,149–156.

Duffey, E., 1964. The terrestrial ecology of Ascension Island. Journal of AppliedEcology 1, 219–251.

Eastwood, A., Gibby, M., Cronk, Q.C.B., 2004. Evolution of St Helena arborescentAstereae (Asteraceae): relationships of the genera Commidendrum andMelanodendron. Botanical Journal of the Linnean Society 144, 69–83.

Farrier, D., Whelan, R., Mooney, C., 2007. Threatened species listing as a trigger forconservation action. Environmental Science and Policy 10, 219–229.

Francisco-Ortega, J., Jansen, R.K., Santos-Guerra, A., 1996. Chloroplast DNA evidenceof colonization, adaptive radiation, and hybridization in the evolution of theMacaronesian flora. Proceedings of the National Academy of Sciences of theUnited States of America 93, 4085–4090.

Francisco-Ortega, J., Santos-Guerra, A., Kim, S.-C., Crawford, D.J., 2000. Plant geneticdiversity in the Canary Islands: a conservation perspective. American Journal ofBotany 87, 909–919.

Frankham, R., 1997. Do island populations have less genetic variation thanmainland populations? Heredity 78, 311–327.

Gray, A., 2003. Aspects of the Ecology of Euphorbia origanoides L. UnpublishedReport to USAF Ascension Island.

Gray, A., 2009. Ascension Spurge Euphorbia origanoides L. Climate and ViabilityStudy: Final Report Unpublished Report to USAF Ascension Island.

Gray, A., Gardner, S., Kirk, L., Robinson, P., Smolka, Z., Webster, L., 2000. The Statusand Distribution of the Endemic Vascular Flora of Ascension Island. Universityof Edinburgh, Edinburgh.

Gray, A., Pelembe, T., Stroud, S., 2005. The conservation of the endemic vascularflora of Ascension Island and threats from alien species. Oryx 39, 449–453.

Harley, J.L., Harley, E.L., 1987. A check-list of mycorrhiza in the British flora. NewPhytologist 105, 1–102.

Hess, G.R., Koch, F.H., Rubino, M.J., Eschelbach, K.A., Drew, C.A., Favreau, J.M., 2006.Comparing the potential effectiveness of conservation planning approaches incentral North Carolina. USA Biological Conservation 128, 358–368.

Heywood, V.H., 1998. Flowering Plants of the World. BT Batsford Ltd., London.Hooker, J.D., 1867. On insular floras. Gardeners’ Chronicle and Agricultural Gazette,

pp. 6–7, 27, 50–51, 75–76.IUCN, I.S.S.G., 2008. The Global Invasive Species Database. <http://www.issg.org/

database/welcome/>.Juan, C., Emerson, B.C., Oromí, P., Hewitt, G.M., 2000. Colonization and

diversification: towards a phylogeographic synthesis for the Canary Islands.Trends in Ecology & Evolution 15, 104–109.

Kew, R.B.G., 2008. Seed Information Database: <http://data.kew.org/sid/SidServlet?Source=epic&ID=32712&Num=L7r> (accessed 11.07.08).

Kim, H.G., Keeley, S.C., Vroom, P.S., Jansen, R.K., 1998. Molecular evidence for anAfrican origin of the Hawaiian endemic Hesperomannia (Asteraceae).Proceedings of the National Academy of Sciences of the United States ofAmerica 95, 15440–15445.

Koske, R.E., Gemma, J.N., Flynn, T., 1992. Mycorrhizae in Hawaiian angiosperms: asurvey with implications for the origin of the native flora. American Journal ofBotany 79, 853–862.

Lawton-Rauh, A., Robichaux, R.H., Purugganan, M.D., 2003. Patterns of nucleotidevariation in homoeologous regulatory genes in the allotetraploid Hawaiiansilversword alliance (Asteraceae). Molecular Ecology 12, 1301–1313.

Lawton-Rauh, A., Robichaux, R.H., Purugganan, M.D., 2007. Diversity and divergencepatterns in regulatory genes suggest differential gene flow in recently derivedspecies of the Hawaiian silversword alliance adaptive radiation (Asteraceae).Molecular Ecology 16, 3995–4013.

Linnaeus, C., 1753. Species Plantarum Imprensis Laurentii Salvii, Holmiae.Lym, R.G., 1998. The biology and integrated management of leafy spurge (Euphorbia

esula) on north Dakota Rangeland. Weed Technology 12, 367–373.Mott, K.A., Gibson, A.C., O’Leary, J.W., 1982. The adaptive significance of

amphistomatic leaves. Plant, Cell and Environment 5, 455–460.Muthukumar, T., Udaiyan, K., 2000. Arbuscular mycorrhizas of plants growing in the

Western Ghats region, Southern India. Mycorrhiza 9, 297–313.Nielson, D.L., Sibbett, B.S., 1996. Geology of Ascension Island, South Atlantic Ocean.

Geothermics 45, 427–448.Noy-Meir, I., 1985. Desert ecosystem structure and function. In: Evenari, M. (Ed.),

Ecosystems of the World 12A: Hot Deserts and Arid Shrublands. Elsevier,Amsterdam, pp. 93–103.

Pearce, F., 2004. The accidental rainforest. New Scientist 2465, 44–45.Raunki�r, C., 1934. The Life Forms of Plants and Statistical Plant Geography, being

the collected papers of C. Raunki�r. Oxford University Press, Oxford.Rivas-Martínez, S., Sánchez-Mata, D., Costa, M., 1999. North American boreal and

western temperate forest vegetation (Syntaxonomical synopsis of the potentialnatural plant communities of North America, II). Itinera Geobotanica 12, 5–316.

RSPB, 2007. Costing Biodiversity Priorities in the UK Overseas Territories, G.C.L.Prepared by Mark Rayment. RSPB London, <www.rspb.org.uk/ourwork/library/reports.asp>.

Salisbury, E.J., 1928. On the causes and ecological significance of stomatalfrequency, with special reference to the Woodland flora. PhilosophicalTransactions of the Royal Society of London Series B 216, 1–65 (Containingpapers of a biological character).

Sang, T., Crawford, D.J., Kim, S.C., Stuessy, T.F., 1994. Radiation of the endemic genusDendroseris (Asteraceae) Centaureinae) based on ITS sequences of nuclearribosomal DNA. Plant Systematics and Evolution 214, 147–160.

Schilling, E.E., Paner, J.L., 2002. A revised classification of subtribe Helianthinae(Asteraceae: Hetiantheae). I. Basal lineages. Botanical Journal of the LinneanSociety 140, 65–76.

Sharma, D., Kapoor, R., Bhatnagar, A.K., 2008. Arbuscular mycorrhizal (AM)technology for the conservation of Curculigo orchioides Gaertn.: anendangered medicinal herb world. Journal of Microbiology and Biotechnology24, 395–400.

Silvertown, J., 2004. The ghost of competition past in the phylogeny of islandendemic plants. Journal of Ecology 92, 168–173.

Simberloff, D., 1998. Flagships, umbrellas, and keystones: is single-speciesmanagement passé in the landscape era? Biological Conservation 83, 247–257.

Smith, W.P., Zollner, P.A., 2005. Sustainable management of wildlife habitat and riskof extinction. Biological Conservation 125, 287–295.

Stace, C., 1997. New Flora of the British Isles. Cambridge University Press,Cambridge.

Stüder, T., 1889. Die Forschungreise der. Gazelle Bd. 1, pt. III. Berlin.Thoen, D., 1987. First observations on the occurrence in hydrophytes, hygrophytes,

halophytes and xerophytes in the region of Lake Retba of vesicular–arbuscularmycorrhizae (VAM) (Cap-Vert, Senegal) during the dry season. Mémoire de laSociété Royale de Botanique de Belgique 9, 60–66.

Trappe, J.M., 1987. Phylogenetic and ecologic aspects of mycotrophy in theangiosperms from an evolutionary standpoint. In: Safir, G.R. (Ed.),

1766 A. Gray et al. / Biological Conservation 142 (2009) 1754–1766

Ecophysiology of VA Mycorrhizal Plants. CRC Press, Boca Raton, Florida, USA, pp.5–25.

UNEP, 1992. World Atlas of Desertification. Edward Arnold, London.Varma, A., Hock, B., 1999. Mycorrhiza: Structure, Function, Molecular Biology and

Biotechnology. Springer-Verlag, Berlin, Germany.Wagner, W., Funk, V., 1995. Hawaiian Biogeography: Evolution on a Hot Spot

Archipelago. Smithsonian Institution Press, Washington DC.Wait, D.A., Aubrey, D.P., Anderson, W.B., 2005. Seabird guano influences on desert

islands: soil chemistry and herbaceous species richness and productivity.Journal of Arid Environments 60, 681–695.

Wang, B., Qiu, Y.L., 2006. Phylogenetic distribution and evolution of mycorrhizas inland plants. Mycorrhiza 16, 299–363.

Wehi, P.M., Clarkson, B.D., 2007. Biological flora of New Zealand 10. Phormium tenax,harakeke, New Zealand flax. New Zealand Journal of Botany 45, 521–544.

Yeo, T., Barker, G., Caton, M., Challen, C., Chaytor, D., Horwood, M., Hurd, M. N.,Lazarowicz, M., Liddell-Grainger, I., Malik, S., Riordan, L., Stuart, G., Swinson, J.,Turner, D., Walley, J., Woolas, P., 2008. House of Commons Environmental AuditCommittee Halting UK biodiversity loss. Stationary Office Limited, London.