<<

Pacific Science (1977), vol. 31, no. 4 © 1978 by The University Press of . All rights reserved

Vegetation and Urbanization on Majuro , !

MICHAEL D. SABATH 2

ABSTRACT: The urban and nonurban vegetation on the Micronesian atoll of Majuro is described, including changes in forest canopy, understory shrubs, yards, and cultural features since urbanization began in 1944. Currently, the nonurban areas are covered with Cocos nucifera (coconut) groves mixed with smaller Artocarpus altilis (breadfruit) groves, which were probably established in the late 1800s. Indigenous vegetation is limited to a narrow band along the ocean or shoreline, or as minor understory species in the Cocos­ Artocarpus groves. A United States military base was established in 1944 on three eastern islands of the atoll (Uliga, Dalap, and ). This has sub­ sequently been developed into a major administrative and commercial center for the Marshall Islands. Urbanization on Uliga, Dalap, and Djarrit has resulted in reduction of tree canopy; establishment of extensive yards with grasses, herbs, and sedges; and reduction of many indigenous and aboriginally introduced understory species. Nevertheless, some aboriginally introduced and indigenous species remain in the urban areas as important species (Cocos, Artocarpus, and Tournefortia), with many being actively propagated. Orna­ mental species, which have expanded in importance, especially in the shrub layer, consist primarily of species recorded in Laura village prior to urbaniza­ tion. The urban plant community is a mixture of indigenous, aboriginally introduced, and recently introduced species. Future urban expansion is predicted with commercial and residential development replacing horticultural forests along the southern islands of the atoll.

MAN HAS AFFECTED the distribution and 41 species on Kapingamarangi (Niering 1963) abundance of plant species on the Pacific and 44 on Arno (Hatheway 1953). The arrival islands from the earliest contacts, but urban­ of ocean-going cultures began a process of ization is a relatively recent and infrequent plant introduction that significantly increased occurrence (Fosberg 1963). Although urban the vascular flora of most islands. For ecosystems are expanding units, descriptions example, aboriginal introductions brought and ecological investigations of their unique 16 species tb' :Kapingamarangi (Niering 1963) biological communities are rare (Schmid and 24 species to Arno (Hatheway 1953). The 1975). The current report describes the urban atoll agriculturalists have removed orreduced and nonurban vegetation of a Pacific atoll the indigenous plant communities, replacing undergoing rapid change since 1944. them with communities dominated by useful The extensive alteration and manipulation species. Only the driest have escaped of plant communities by indigenous agricul­ major vegetation changes (Fosberg 1956). turalists on Pacific atolls has been commented Trading and missionary influence began in upon frequently (Bates and Abbott 1958, the 1800s followed by military operations in Niering 1963). On the atolls, the indigenous the 1900s imposing further changes on the vascular flora is limited, with for example, atoll ecosystems. In the early 1900s the commercial and administrative centers of European and Japanese interests became I Manuscript accepted 19 July 1977. 2 School ofAustralian Environmental Studies, Griffith established on various of the atolls. University, Nathan, 4111, . With the onset of World War II, Japanese 321 MAJURO A

2 o 2 4 6

KILOMETRES ...pi~.,,\ ROQuron /1/ ':;'~\(~;:~i'\]' .. i<.'>;

: ; .../: / 07°10'''' I / f

j / ('''i/:': Loura or \ ~. \f Majuro Village ...... ,...... ~ \...... \ /~::::) ···....-.._"... .-'" /1 ...... - } l

Majuro

..•...... , , , ...... "' .....

Ajilt

FIGURE I. Map of Majuro Atoll, Marshall Islands.

!I!ft€E . 5 $_ 4· 1··Mb! -&5 8M; m .6" , PACIFIC

...... ~~ . .: ---_::..::::...-.~::::...-..:..- __...---'--::...... :...---:...... -~~~ ., ..•

HTIaklJ '. ~~.:::,::::;;;::t:::::'::~\;;~"

-~~.- ...... ~..... Arniel ,.:.!.!.~~~~~::;::~·.ti·········· ...\Djarrlt

Bokelljker BOkelatule ...... ······.....~~iga

\..... nlllgere

....)Dalap orcos' Ijemwa .=JR~.'!~!!.~~~.!.~ ~..r#fi!I!f.~!!i!iIiiiiii_~~~~(i··~···~···'!'!i··~···~· ..~..:::r~~-t:. ~ !!..~p.!!'!'••••. "".\ ;: -::: " _.... . E"'," :::::~.... " Anenelibw \ Lokojbar .lIneiokren 171°15' 324 PACIFIC SCIENCE, Volume 31, October 1977 and then United States military bases were taken in 1970 from Arniel along the eastern established and abandoned. A few centers and southern island chain to Majuro Island have continued in importance. Majuro and at 1800 to 2500 ft altitude was used to con­ Kwajalein, Marshall Islands, have been devel­ struct detailed maps of vegetation and oped for commercial, administrative, and cultural features on Arniel, Rairik, Uliga, military purposes, leading to population and Majuro. The locations and areas covered concentration and urbanization (Spoehr by buildings, roads, and cleared areas were 1949). determined from the photographs and major characteristics were verified on the ground. A tree survey was completed on sections of Arniel, Uliga, Rairik, and Majuro by METHODS AND MATERIALS locating and identifying each tree in the 1970 Majuro Atoll (7°07' N, 171°12' E) is aerial photographs on the ground. A tree was located in the wet zone ofthe southern defined as any plant with greater than 10 cm chain of the Marshall Islands. The atoll is diameter at breast height, or for Pandanus about 297 km2 in area, with 64 islands tectorius Soland and Cocos nucifera L., any totaling 9.17 km2 of dry land (Bryan 1971; specimen greater than 2 meters in height. see Figure 1 for map). Both temperature Since the Cocos could be readily identified (annual average 27°C) and rainfall (annual and located on the photographs without average 308 cm) are relatively uniformly ground verification, only those individuals distributed throughout the year (Wiens 1962). that had died since 1970 were recorded. From Prior to 1944 the atoll economy was based the photographs the following could be esti­ upon subsistence horticulture and mated: density of each species; foliage cover export (Spoehr 1949). Urbanization com­ of major species; and individual location menced in 1944 with the United States relative to cultural structures or saltwater. military occupation of Dalap, Uliga, and In addition, aerial photography (oblique Djarrit in the eastern end of the atoll. At at 1000 feet altitude and vertical at an present, Majuro is being used for traditional undetermined altitude) completed in January horticulture, for copra production, and as and December 1944, respectively, provided an urban center. some evidence for change in vegetation The atoll was visited in 1974. Vegetational patterns and cultural development over the and cultural features were described quan­ 30 years since urbanization began. titatively on the islets ofArniel, Uliga, Rairik, In 1974, the shrub layer was analyzed on and Majuro. Qualitative descriptions were Uliga, Rairik, and Majuro to determine completed on Eniligere, Biggariat, Utwe, patterns of association with dwellings. A Enirak, Calalin, Allokan, Anemakij, shrub was defined as any specimen less than Bokelijlker, Bokelatule, Lejemwa, and 10 cm diameter at breast height, or for Limeiokren (place names according to Bryan Pandanus and Cocos, any specimen less than 1971; see Figure 1). The locations for plant 2 meters in height. On Rairik and Majuro community analysis were determined by the four transects were laid out at right angles availability of transport and aerial photog­ from each house site. Each transect consisted raphy. Arniel was chosen as a relatively of a volume 70 meters in length, 1 meter in isolated island, without permanent occupants width, and 1 meter in height beginning 1 and reached only by boat. Rairik and the meter from ground level. All shrubs inter­ Ajiltake region of Majuro Island had per­ secting the transect volume were counted manent residents and were accessible by a according to species and distance from the main roadway, but contained forested areas central dwelling. Only house sites that had still used for horticulture. Uliga was one of been used for more than 10 years were the three major urbanized islets with govern­ included in the survey. A total of 25 house ment, commercial, and residential develop­ sites were sampled on both Majuro and ment. Rairik. On Uliga the dwelling density pre­ A vertical aerial photographic sequence vented using this sampling technique on each Vegetation and Urbanization on Majuro-SABATH 325

house. A crude shrub-layer description was military complex which established the major derived from three cross-island transects. roads, runways, and building sites that existed in 1974. Cocos and Artocarpus had been extensively cleared, but on Uliga there RESULTS seemed to be less activity. Urbanization Since 1944 Unfortunately photographs of the Ajiltake region of Majuro were not available, but few In the oblique aerial photo­ families lived in the area in 1947 (Spoehr graphs, Rairik appeared undisturbed by 1949). The roadway connecting Majuro with development and without indication of resi­ the eastern islets was built in 1961 (Kabua dences along the lagoon. By December 1944 and Pollock 1967) and undoubtedly contrib­ the vertical aerial photographs reveal areas uted to further occupation of the Ajiltake on Rairik being cleared of vegetation and an region. airplane runway established on the eastern Based upon the 1944 photographs, the end. The Rairik study area was undisturbed eastern and southern chain of islets were except for a small section that has since been covered with a mixed Cocos-Artocarpus planted in Cocos. The major Artocarpus altilis grove prior to development. From the history (Parkinson) Fosberg, groves were evident, of development in the Marshall Islands but the photographic scale does not allow (e.g., Arno) the extensive Cocos groves were identification of individual trees. During probably established over much of the atoll 1944 the roadway that is still in use, con­ in the late 1800s and early 1900s (Hatheway necting Rairik with Dalap, was constructed. 1953). In 1944 all Cocos groves appeared Many house sites were evident in the same mature, and are described currently by local positions as presently occupied. informants as dating prior to World War II. Photographs of Arniel in January 1944 In 1945 the military demobilized rapidly, showed the general pattern of Artocarpus and leaving only a small administrative unit by Cocos groves similar to that in 1974. How­ 1947. The Marshallese population began ever, the large Calophyllum inophyllum L. reoccupation of Djarrit and Dalap islands were not in evidence on the lagoon shore in (Spoehr 1949). Subsequently, the Dalap­ 1944. House sites were in the same location Uliga-Djarrit area became the administrative in both years. The December 1944 photo­ and commercial center for the atoll and for graphs indicated no further development on the Marshall Islands. The atoll population Arnie!. increased rapidly through migration and Excellent oblique photographs taken in natality with the Djarrit-Dalap-Uliga area January 1944 exist for Djarrit, but Uliga and receiving the greatest increase (Table I). part of Dalap were photographed only at Uliga, Eniligere, Biggariat, and Djarrit are some distance. Djarrit was occupied by a currently urbanized with dense populations small Japanese naval base beginning in 1940, in mixed residential, commercial, and govern­ but there were less than a few dozen buildings. ment land use. Both Dalap and Djarrit were occupied by Marshallese prior to World War II, and at Comparisons among Islands, 1974: Vegetation the beginning of 1944 there were small Cover villages of about 400 people (Spoehr 1949). Uliga was not reported as being occupied in The relative proportion of forest canopy, 1944 by either Japanese or Marshallese, and open yards, roadways, and buildings as the photographs do not indicate any develop­ determined from the 1970 aerial photographs ment. The island appeared to be covered is presented in Table 2. Majuro, Rairik, and with Cocos and possibly with some Arniel were forested islands with Cocos as a Artocarpus. dominant. The Cocos canopy was occasion­ By December 1944 the vertical photo­ ally broken by Cordia subcordata Lam., graphs confirm that the Djarrit-Uliga-Dalap Guettarda speciosa L., Artocarpus, Hernandia area was occupied by a large United States sonora L., Ficus tinctoria Forst. f., and

.'.~.~.. '" 326 PACIFIC SCIENCE, Volume 31, October 1977

TABLE 1 POPULATION HISTORY OF MAJURO

POPULATION

DATE Laura* DUDt Total SOURCE

1935 785 Bryan 1971 1948 1,236 t Bryan 1971 1958 1,028 2,387 3,415 High Commissioners Office, 1959 1970 7,165 Bryan 1971 ** 1970 1,572 5,829 7,301 U.S. Department of Commerce, Bureau of the Census 1971 ** 1973 2,287 8,003 10,290 Office of the High Commissioner, 1974

·Laura is the village on Majuro Island. t DUD is the Djarrit-Uliga-Dalap region, now loosely referred to as Rita. : SpaehT (1949) reported that most of the 1947-1948 population was living on Majuro Island (Laura village). "The differences in total population reported by Bryan and the Bureau of the Census in 1970 cannot be resolved, but the discrepancy does not alter the overall pattern of population change.

TABLE 2 tectorius) usually only a few individuals in PROPORTION OF VEGETATIONAL AND CULTURAL width occupied the ocean and lagoon shore­ FEATURES IN THE STUDY AREAS ON MAJURO line. Arniel contained a small Ochrosia oppositif'olia (Lam.) K. Schum. grove (23 CATEGORY MAJURO RAIRIK ARNIEL ULiGA individuals) on a boulder rampart facing the Housing 0.0012* 0.0038 0.0016 0.2057 ocean. Cocos groves occupied the remaining Roads 0.0425 0.0422 0.0000 0.1759 areas of the three islands. Yards 0.0040 0.0000 0.0091 0.5587 A single macadam road connected the Forest canopy 0.9523 0.9540 0.9893 0.0597 administrative center in the Djarrit-Uliga­ Dalap (Rita) area with the older village NOTE: Housing was considered to be buildings or residential dwellings. Roads were roadways actually in use and without vegetation. Yards were center (Laura) on Majuro Island. The road­ areas cleared of trees and shrubs, but with a ground layer of grasses. herbs. way through the length ofRairik and Majuro and sedges. Forest canopy was considered to be areas with tree cover over 2 meters in height. islands formed a minor part of total land use. • Proportion of coverage. Arniel had a narrow path along the lagoon but was without roads. There were scattered Calophyllum inophyllum L. Smaller individ­ houses on Arniel, Majuro, and Rairik that uals of these species, along with M orinda represented minor land use. Houses were citrifolia L., Pipturus argenteus (Forst. f.) constructed on a shallow coral rubble bed Wedd., Premna obtusif'olia L., Pandanus which was largely cleared of grasses, herbs, tectorius, Terminalia sp., and young Cocos and most common understory shrubs for nucifera formed the understory. The Arto­ several meters from the dwellings. The small carpus groves contained Cocos, Hernandia, yards were planted with some ornamental and Cordia, with understory species similar and food-producing trees and shrubs. On to the Cocos groves. Scattered Artocarpus Rairik the tree canopy enclosed the houses, groves occupied the interior sections of each reducing open yards to a minimum. The island. The extent of Artocarpus groves was combined yards, houses, and roadways con­ variable: on Rairik 26 percent of the total stituted less than 5 percent of the land area area; on Arniel 20 percent of the total area; mapped on the three islands. and on Majuro 10 percent of the total area. Urbanization had significantly reduced the A narrow band of indigenous species forest canopy on Uliga (only 6 percent of (Guettarda speciosa, Scaevola tace'ada L., total area), substituting a dominance of open Tournef'ortia argentea L.f., and Pandanus yard (56 percent) with grasses, herbs, and Vegetation and Urbanization on Majuro-SABATH 327

TABLE 3

DENSITIES OF TREE SPECIES ON MAJURO, RAIRIK, ARNIEL, AND ULIGA

SPECIES MAJURO ULIGA

Indigenous Pandanus tectorius* 23.50 2.70 Guettarda speciosa 2.95 0.11 Scaevo/a taccada 1.73 0.23 Cordia subcordata 0.10 0.04 Tournefortia argentea 0.92 4.28 Pemphis acidu/a 0.20 Pipturus argenteus 0.71 Premna obtusifo/ia Hernandia sonora Bruguiera conjugata Tennina/ia sp. 0.08 Ochrosia oppositifo/ia 0.04 Aboriginal Cocos nucifera 120.68 22.26 Artocarpus a/ti/is 4.27 0.45 Ca/ophy//um inophy//um 0.81 1.09 Ficus tinctorius Morinda citrifo/ia Recent Musa sp. 0.41 0.64 Carica papaya 0.20 0.60 P/umeria 2 sp. 1.58 Erythrina variegata 0.79 Hibiscus sp., ornamental 0.04 Casuarina sp. 0.19 Ca/otropis gigantea 0.11

NOTE: Origins are based upon Niering (1963), Hatheway (1953), and St. John (1951). *Pandanus should be included as either indigenous or aboriginally introduced, depending upon cultivar. sedges surrounding houses (21 percent). exception was Tournefortia which showed Houses were associated with a shrub layer marked increase in density on Uliga. The dominated by ornamental species. Road­ increase might be explained as follows: ways, paved or dirt, occupied 18 percent of Tournefortia occurred commonly on all the total area. islands as a shrub or small tree along lagoon or ocean shorelines. The clearing for yards on Uliga reduced the shoreline vegetation Comparisons among Islands, 1974: Tree Dis­ layer including Tournefortia, but most of the tribution remaining individuals were tree size. Perhaps In Table 3 the tree densities (as individuals the selective clearing has reduced light and per hectare) on Majuro, Rairik, Arniel, and nutrient competition thereby increasing the Uliga are grouped according to introduction density of tree-size Tournefortia. history. Pandanus was represented by indig­ The aboriginally introduced Cocos and enous as well as introduced cultivars which Artocarpus were the most common canopy are widely propagated for food. On Arno trees on Majuro, Rairik, and Arniel, but active propagation of Bruguiera conjugata were si¥nificantly reduced on Uliga. Ficus (L.) Herr. led Hatheway (1953) to suggest was a rare tree in Artocarpus groves' or in that the species might not be indigenous. close association with houses. Morinda rarely With one exception, the density of indig­ attained sufficient size to be classified as a enous species on Uliga was reduced in tree, but is widespread as an understory comparison with nonurbanized islands. The shrub in Cocos groves. Thus, the importance 328 PACIFIC SCIENCE, Volume 31, October 1977

,j ,I !

\ PondDnus 'llctor/US , 20 ~ ,,

______-9r niel __ Majuro ___ Roirlk .•...... Ullga

______Arniel __ Majuro __ Rairlk ...... • UIIga

......

'" DISTANCE'" '" 10

o+-=o£,-~'>O<::'''::;O''''-.L..-->l...-L-...... ''---~~~--I FIGURE 3. Relative frequency of tree species in rela­ o 10 2030 40 00 eo 1'0 '" '00 L~~ ~STANcr 00000 tion to distance from lagoon or ocean.

FIGURE 2. Relative frequency of tree species in rela­ tion to distance from lagoon or ocean. occur in the study area. Casuarina sp. and Calotropis gigantea (L.) R. Br. were cultivated attached to the presence or absence of tree­ primarily in the urban areas. Carica papaya size Morinda or Ficus on any island is L. was frequently noted near houses on equivocal. Calophyllum was a common large Majuro and Rairik but more frequently on tree along the lagoon shoreline. Houses were Uliga. Musa sp. occurred only on Majuro frequently built adjacent to large specimens and Uliga in close association with houses. that provided shade for work or play. The Groundwater, salinity, ocean spray, and oblique aerial photographs of Arniel and cultivation practices are important determi­ Djarrit in 1944 did not show these easily nants ofplant distribution on a lagoon-ocean recognized trees although they were common gradient (Hatheway 1953, Niering 1963, along the lagoon shoreline in 1974. The high Fosberg 1956). The distributions of the seven density now seen on Uliga may be attributed most common tree species on a cross-island to recent propagation. gradient is presented in Figures 2 and 3 for The recently introduced tree species Majuro, Rairik, Arniel, and Uliga. The showed higher densities on Uliga than any distance of an individual tree was measured other island area. Plumeria sp. and Hibiscus from the lagoon and ocean shorelines, and sp. were common in the immediate vicinity then standardized to a scale of 0 to 100. The of houses on Rairik and Majuro but did not relative abundance of each species is plotted attain tree size. Erythrina variegata (L.) was against the standardized scale. observed occasionally in the secondary forest The trees demonstrated three distinct pat­ along the old World War II runways and terns of distribution. First, a relatively military areas on eastern Rairik but did not uniform distribution was shown by Cocos on

. &i&9&£M4tiMMNt#! iM. Vegetation and Urbanization on Majuro-SABATH 329

all islands. On Uliga there was a slight decrease in Cocos frequencies between 10 and 35 on the scaled gradient, which was associated with a major roadway and building development. Second, an increased abundance on the lagoon and/or ocean shoreline is shown by Guettarda, Tournefortia, Pandanus, Scaevola, and Calophyllum. Pandanus was distributed across all islands with increased abundance I at both lagoon and ocean shores. Scaevola, MOrindo cllrlfolia -----~ .... Pa~anus t«/OOU3 Tournefortia, and Guettarda showed greater .\ ~..: ~=ugg:::;folf'q \ ...•.. 'GlJ.lfardo I/HC;O$Q abundance along the ocean shoreline and +---==;==--r.:...... ,r-""'T--..,,---,;:--:o=-...,,---,;-~ were rarely found in the central areas. On the ocean side, these species along with Pandanus formed a tree-shrub layer that was left FIGURE 4. Relative frequency of ten common shrubs uncleared by the Marshallese and served as in relation to isolated houses. a barrier to salt spray (Hatheway 1953). Calophyllum was most common on the lagoon quency with distance from the dwelling but shoreline with an occasional specimen near were not as restricted in distribution. In the ocean or in the central area. On Uliga, addition to the species frequent enough to be where large-scale clearing had occurred, plotted, occasional specimens of Calotropis similar patterns of species distributions were gigantea, Acalypha grandis Benth., Codiaeum clearly recognizable even though densities of variegatum (L.) Bl., Nerium oleander L., most species were greatly reduced. Hibiscus sp., and Citrus aurantifolia Third, Artocarpus was common in the (Christm.) Swingle were noted within 6 to center of all islands with an occasional 10 meters of house sites. individual near the lagoon on Arniel and Since the immediate house sites were Rairik. The distribution pattern on Uliga usually kept clear of large trees and shade, was similar to that on Rairik, Majuro, and Artocarpus would be expected to be cultivated Arniel but with reduced densities due to at some distance from house sites. Indeed, clearing and development. Artocarpus has a young Artocarpus increased in abundance low tolerance for salt spray and saline between 28 and 42 meters from dwelling sites groundwater, thus the species is restricted to but then declined. These distances may propagation in the centers of larger islands represent a Marshallese optimum distance (Fosberg 1949). for Artocarpus in relation to houses. How­ Figure 2 presents the distribution patterns ever, the adult Artocarpus did not show a for two recently introduced species found as similar relationship with house sites. trees only on Uliga, Erythrina and Plumeria. Premna, Pandanus, Morinda, Guettarda, Both species increased in frequency in the and sprouting Cocos formed the understory island center since large specimens of both layer in Cocos groves. The understory layer species were associated with residences and was cleared from the immediate vicinity of these tended to be located centrally. houses, and the clearing activities are reflected in Figure 4. In order to facilitate copra collection, the shrub layer was slashed from Comparisons among Islands, 1974: Shrubs the Cocos groves once or twice a year. During Shrubs closely associated with houses on the clearing activities some specimens of Rairik and Majuro are plotted in Figure 4. Morinda, Pandanus, Premna, and Guettarda The ornamental Stenolobium stans L. and were left intact since each species is con­ Plumeria sp. were clumped near houses. sidered useful by the Marshallese. Occasional Musa and Carica rapidly declined in fre- specimens of Hernandia, Pipturus, Calo- 330 PACIFIC SCIENCE, Volume 31, October 1977 phyllum, Tournefortia, and Scaevola were The early Marshallese established the also encountered. horticultural subsistence agriculture that Although comparable shrub sampling on persists on many atolls. Cocos nucifera, Uliga could not be completed, the cross­ Artocarpus altitis, Morinda citrifotia, Calo­ island transects revealed some differences. phyllum inophyllum, and numerous cultivars Morinda, Premna, and Pipturus usually asso­ of Pandanus were introduced to the interior ciated with Cocos groves were absent from forests. The areas covered by these aborigin­ Uliga. Crinum sp., Hibiscus, and Codiaeum ally introduced species were relatively minor were used extensively on Uliga as ornamen­ until the Marshallese came under the increas­ tals and occurred in markedly increased ing influence of missionaries and traders. frequency. Young Tournefortia, Scaevola, Artocarpus and Cocos groves were signifi­ Pandanus, and Artocarpus were present in cantly expanded in the late 1800s. In response limited numbers. Several ornamental species to the copra trade, Cocos were planted over such as Pseuderanthemum sp., Delonix regia large areas to become the dominant atoll L., and Cordyline sp. were observed only on tree. For example, in 1952, 69 percent of Uliga. In general, nonornamentals were was planted in Cocos (Hatheway reduced, but a few food species such as 1953). Carica and Musa were commonly cultivated. With the arrival ofmissionaries and traders in the late 1800s a new series of plant intro­ ductions began which brought Erythrina, Additions to Majuro Flora Carica, Citrus, Musa, Pemphis, and Plumeria. Records of general floral collecting on Recent introductions have been dominated Majuro Atoll are limited to a brief visit to by ornamental trees and shrubs. On Arno, Majuro Island resulting in 70 species (St. 20 of the total 48 trees and shrubs were John 1951). Although limited in scope, the recent introductions (Hatheway 1953), with current study revealed seven taxa previously 15 ornamental species. On Majuro, II of 14 unrecorded: Erythrina variegata, Casuarina recently introduced trees and shrubs were sp., Delonix regia, Cordyline sp., Calotropis ornamentals. gigantea, Pluchea odorata Nat. (L.) Cass., Fosberg (1949, 1956) suggested that the and Nerium oleander. Each species was closely wetter Marshallese atolls such as Majuro associated with urban development and un­ would have similar vegetation histories, doubtedly represented recent introductions. species compositions, and associations. Spoehr (1949) describes a sequence of mis­ sionary, trader, and external governmental control for Majuro similar to that of Arno. DISCUSSION Thus, Majuro was probably converted to Hatheway (1953) reconstructed the Cocos plantations beginning in the late 1800s. sequence of vegetation alteration for Arno, By 1944 (based upon aerial photography), an atoll east of Majuro. Prior to human the Cocos groves were established in patterns occupation, the islands were bordered with similar to that seen in 1974 on Majuro, a shrub and tree layer of Scaevola taccada, Rairik, Enirak, Utwe, and the islets from Tournefortia argentea, Pandanus tectorius, Calalin to Djarrit. Terminalia samoensis L., and Guettarda In 1974, the indigenous species were ex­ speciosa. Island interiors were forested with tremely limited in distribution and relative Barringtonia asiatica, Hernandia sonora, abundance on the Co cos- and Artocarpus­ Ochrosia opposit(folia, /ntsia bijuga, Pandanus dominated islets. Lagoon and ocean shoreline tectorius, Guettarda speciosa, Pisonia grandis species (Tournefortia, Scaevola, Pandanus, R. Brown, and Cordia subcordata. On sandy and Guettarda) seemed most successful since interior soils trees such as Allophylus they were retained by Marshallese for salt­ timorensis, Pipturus argenteus, and Premna spray protection. These species persisted even obtusifotia occurred. Pemphis acidula Forst. on the urbanized islets. The cultivation of occupied the saline flats. Cocos and Artocarpus has largely eliminated

'S $ 4$#' 4; .<.: %W"?9 ,*.- i it! '9 .u :&&Z2EU; .-E&i,nO ¥ . Vegetation and Urbanization on Majuro-SABATH 331

TABLE 4 tation, thus precluding direct comparison

POPULATION DENSITIES ON ARNO AND MAJURO, with Arno or Kapingamarangi. 1935-1970 Over the last 30 years, urbanization has altered indigenous and aboriginally intro­ MAJURO ARNO duced plant communities primarily by re­ 2 2 YEAR (lndividuals/km ) (lndividuals/km ) placing forested areas with open yards. The 1935 85.6 73.1 urban vegetation is characterized by scattered 1948 134.8 85.5 trees resulting in open canopy; a high per­ 1970 798.2 53.8 centage of ornamental shrubs; and domi­ nance ofyards containing grasses, herbs, and SOURCE: Based upon Bryan (1971). sedges. Most of the ornamental trees and shrubs had been recorded in Laura village on indigenous forest associations. Ochrosia was Majuro Island prior to major development limited to the small groves on Arniel and (St. John 1951). The recently introduced Anemakij. Pisonia was observed only as a species could be found around the isolated small grove (21 individuals) of understory family dwellings in the Cocos and Artocarpus saplings on Anemakij. Barringtonia asiatica groves on Rairik and Majuro. Indeed, the was absent as a tree, but was being cultivated urban shrub and tree communities appeared by the Agriculture Department for orna­ to be an extension of patterns set in these mental plantings. isolated dwellings. If rural dwellings with Regeneration ofindigenous forest was pos­ their closely associated plants were placed sibly occurring on eastem Rairik island on together at an urban density, then the existing abandoned military areas. Secondary growth urban vegetation patterns could easily be contained indigenous Terminalia, Hibiscus derived. tiliaceus, Tournefortia, Guettarda, and Although the urban areas contained a high Scaevola, but also introduced species such percentage of ornamentals, indigenous and as Erythrina and Cocos. Although the north­ aboriginally introduced species were also western islands near Roguron were not common. For example, the trees were domi­ observed, it seems unlikely that Majuro has nated by Cocos and Tournefortia. Calo­ any significant areas of indigenous forest. phyllum and Pandanus were common shade For comparison, Arno had 18 percent of the trees. A number of new plantings of Guet­ total area covered by some indigenous forest tarda, Scaevola, and Tournefortia were type (Hatheway 1953). The greater extent of evident, but understory species such as Cocos and Artocarpus on Majuro may reflect Morinda, Pipturus, and Premna were absent. the greater population density evident since The establishment in 1944 of a major 1935 (see Table 4), placing greater demands military center on Majuro followed by upon the horticultural subsistence agriculture governmental and commercial activities and copra production. stimulated urbanization resulting in rapid The plant communities on nonurbanized population growth (see Table 1). The future islands of Majuro Atoll closely resembled expansion of urbanized areas seems assured. those described for other Pacific atolls Since 1970 a new airfield to accommodate (Fosberg 1949, 1956). Although quantitative major jet air traffic has been built over description is available for Arno (Hatheway Anenelibw and Lokojbar islets (see Figure 1953) and the East Carolinian atoll of 1). The mile-long runway with adjacent Kapingamarangi (Niering 1963), direct com­ roadway and passenger-freight processing parison with Majuro is not easily accom­ areas established another major urban vege­ plished. Both descriptions rely on data tation center. In 1974, Tournefortia, Scaevola, determined from "p].1re" stands ofArtocarpus and Cocos were beginning to reinvade the or Cocos. The difficulties of a typological construction areas, and ornamental plantings approach to plant communities are well were commencing. The airport location, known (Kershaw 1973). The current study approximately midway between the Djarrit­ was completed on islands with mixed vege- Uliga-Dalap area and Majuro , seems 332 PACIFIC SCIENCE, Volume 31, October 1977 likely to stimulate further development on on Majuro. The project could not have been both Rairik and the eastern end of Majuro completed without the extensive field as­ islet. New residential construction was pro­ sistance provided by Ritten Nathan. The ceeding on Rairik with several Gilbertese Marsh Fund of the National Academy of families recently occupying new house sites. Sciences (U.S.) and the Society of Sigma Xi Speculation among Marshallese regarding supplied financial aid for field expenses. continued tourist development on both Aubrey Chandica of Griffith University Rairik and Majuro was evident. introduced me to some finer points of inter­ The population increase with subsequent preting aerial photography, completed maps urban development has resulted from a shift and illustrations for analysis, and provided in economic base from subsistence horticul­ the figures in the text. Ruth Forrest and ture to an import-export monetary economy. Janine Smith provided excellent typing as­ Although the final dimensions of the urban­ sistance. ization cannot be predicted at present, certainly the expanding emphasis on Majuro as a governmental and commercial center LITERATURE CITED will probably widen urban development to include all of Rairik and the Ajiltake region BATES, M., and D. ABBOT. 1958. Coral island of Majuro islet. Some residential and com­ -portrait ofan atoll. Charles Scribner and mercial construction was continuing on Sons, New York. 254 pp. Dalap in small sections of uncleared land. BRYAN, E. H., Jr. 1971. Guide to place names The lack of roadways along the northern in the Trust Territory of the Pacific Islands. island chain beyond Dalap seems, for the Pacific Scientific Information Centre, present, to be a barrier to urban development. Bernice P. Bishop Museum, . Unfortunately, comparison with other FOSBERG, F. R. 1949. Atoll vegetation and Pacific urban areas is not possible since salinity. Pac. Sci. 3: 89-92. detailed descriptions are unavailable. While --. 1956. Military geography of the considerable information has been obtained Northern Marshalls. Intelligence Division, on the floristics of many islands, few data Headquarters U.S. Armed Forces Far are available on the structure of the plant East, Tokyo. 320 pp. communities. Based upon observations on ---. 1963. Man's place in the island Majuro it would appear that a distinct, new ecosystem. A symposium. Bishop Museum man-controlled plant community has become Press, Honolulu. 264 pp. established in the Pacific islands. The com­ HATHEWAY, W. H. 1953. The land vegetation ponents represent indigenous, as well as ofArno Atoll, Marshall Islands. Atoll Res. aboriginally and recently introduced species. Bull. 16: 1-68. Some species are successfully sexually re­ HIGH COMMISSIONERS OFFICE, GUAM. 1959. producing while others are simply clones. Census Rept., Census of the Trust Terr. of The dynamic interaction of climate, soils, the Pac. Is. 1958. the component species, and man to form the KABUA, P., and N. POLLOCK. 1967. The new community remain largely unexplored. ecological bases of political power in Laura community. Pages 1-83 in L. Mason, ed. The Laura report. Department ACKNOWLEDGMENTS of Anthropology, University of Hawaii, Honolulu. The following people have aided the pro­ KERSHAW, K. A. 1973. Quantitative and ject by offering information or advice: B. dynamic plant ecology. E. Arnold. 2nd ed. Melde, D. Abel, E. H. Bryan, J. S. Ravina, London. 308 pp. D. Anderson, P. Van Royen, B. Bender, and NIERING, W. A. 1963. Terrestrial ecology of C. Lamoureux. Liki Labi extended aid and Kapingamarangi Atoll, Caroline Islands. the support of the Agriculture Department Ecol. Monogr. 33: 131-160.

it • i 1 NAt' _*i?SW. Iii tM1i£JZ22£ Vegetation and Urbanization on Majuro-SABATH 333

OFFICE OF THE HIGH COMMISSIONER, SAIPAN. SPOEHR, A. 1949. Majuro, a village in the 1974. Population tables. Trust Terr. of the Marshall Islands. Fieldiana: Anthropol. Pac. Is. September 1973 Census. High­ 39: 1-266. lights. U.S. DEPARTMENT OF COMMERCE, BUREAU OF ST. JOHN, H. 1951. Plant records from Aur THE CENSUS. 1971. Number of inhabitants. Atoll and Majuro Atoll, Marshall Islands, Census of the Population, Trust Terr. of . Pacific plant studies 9. Pac. the Pac. Is. 1970. Sci. 5: 279-286. WIENS, H. J. 1962. Atoll environment and SCHMID, J. A. 1975. Urban vegetation. De­ ecology. Yale Univ. Press, New Haven. partment of Geography, University of 532 pp. Chicago, Res. Paper 161, Chicago.