SUMMARY REPORT ON FORESTS OF THE MATAQALI
NADICAKE KILAKA, KUBULAU DISTRICT, BUA, VANUA LEVU
By Gunnar Keppel (Biology Department, University of the South Pacific)
I was approached by Dr. David Olson of the Wildlife Conservation Society (WCS) to assess the type,
status and quality of the forest in Kubulau District, Bua, Vanua Levu. I initially spent 2 days, Friday
(28/10/2005) afternoon and the whole of Saturday (29/10/2005), in Kubulau district. This invitation
was the result of interest by some landowning family clans (mataqali) to protect part of their land and
the offer by WCS to assist in reserving part of their land for conservation purposes. On Friday I visited
two forest patches (one logged about 40 years ago and another old-growth) near the coast and Saturday
walking through the forests in the center of the district.
Because of the scarcity of data obtained (and because the forest appeared suitable for my PhD
research), I decided to return to the district for a more detailed survey of the northernmost forests of
Kubulau district from Saturday (12/11/2005) to Tuesday (22/11/2005). Upon returning, I found out that
the mataqali Nadicake Nadi had abandoned plans to set up a reserve and initiated steps to log their
forests. Therefore, I decided to focus my research on the land of the mataqali Nadicake Kilaka only.
My objectives were the following:
to determine the types of vegetation present
to produce a checklist of the flora and, through this list, identify rare and threatened species in
to undertake a quantitative survey of the northernmost forests (lowland tropical rain forest) by
setting up 4 permanent 50 ×50m plots
to assess the status of the forests
to determine the state and suitability of the proposed reserve
to assess possible threats to the proposed reserve.
to survey the thoughts and feelings of the community of Kilaka Village regarding the WCS
project and setting aside an area as a reserve
During both visits I conducted a reconnaissance by walking through the forest, identifying dominant
species and collecting species in flower or fruit. I also identified as many species as possible (some
were only identified to generic level only) and obtained local names wherever possible. During my
second visit I also set up four permanent 50 ×50m plots (see Fig. 1 for locations) marked with PVC
pipes in lowland tropical rainforest of the northernmost corner of the intended reserve, in which I
measured and identified every tree of 10cm or more in diameter.
Perceptions of the Kilaka Village community regarding the assistance provided by WCS and the idea
of setting aside of forest for a reserve were determined by having formal and informal discussions
while drinking kava. Although kava sessions unfortunately generally exclude young and female
members of the community, it allows detailed discussions with male youths and older male members of
the community, who are the group making discussions regarding land issues.
There are four primary vegetation types in Kubulau district; coastal vegetation (not surveyed), wetlands
dominated by Pandanus tectorius
(not surveyed), mesic forests and rain forest. Before commenting on
the two latter forest types, I would like to draw attention to the presence of extensive wetlands, which
are highly threatened in Fiji (Ash 1992). Although these have been highly impacted by local residents
through draining and burning, restoration efforts could be worthwhile, considering their rarity in Fiji
and that most of the remaining wetlands are severely disturbed.
In addition to these primary vegetation types, secondary types that were created anthropogenically,
including plantations and gardens
, pasture (grasslands maintained by cattle grazing), talasiga
(grasslands maintained by burning) and secondary forest (that is at various stages of recovery after
logging), are present. My survey focuses on “old-growth” forests and, therefore, the remainder of this
section will focus on mesic and rain forests.
Mesic forest is the dominant forest type in the low-lying, near coastal areas of Kubulau and stretches
along a rainfall gradient of lower rainfall near the coast and increasing rainfall inland. It hence provides
a unique opportunity to study changes in climate and plant composition along a rainfall gradient.
However, this forest has also suffered the most from logging, especially on the ridges, where
and Agathis vitiensis
are most abundant. Towards the coast Intsia bijuga
) is a
dominant component of this forest, as are Fagraea gracillipes (buabua), Myristica castaneifolia
(male), Kingiodendron platycarpum (cibicibi), Maniltoa sp. (cibicibi levu), Cynometra insularis
(cibicibi lailai), Gymnostoma vitiense (cau) and Dacrydium nidulum (yaka). Further inland (probably
associated with greater rainfall) Intsia bijuga, Fagraea berteroana and Kingiodendron platycarpum
decrease in importance, and Gymnostoma vitiense and Dacrydium nidulum become increasingly
restricted to ridges, while Endospermum robbieanum (vulavula), Heritiera ornithocephala (savai),
Agathis macrophylla (tahua mahadre), Sterculia vitiense (waciwaci) and Podocarpus neriifolius
(bauwaka) increase in importance.
This is mostly found in the northern portion of Kubulau District at elevations of 200 m or more. Much
of it is located on flat or gently sloping terrain but there are also steeper slopes. Dominant components
of this forest are Retrophyllum vitiense (tahua salusalu), Calophyllum spp. (damanu), Myristica spp.
(male, malamala), Gironniera celtidifolia (masivau), Parinari insularum (sea), Semecarpus vitiensis
(malawaci), Pagiantha thurstonii (tabua kei ra kalavo) and Syzygium spp. (yasiyasi). Other important
species include Palaquium porphyreum (bauvudi), Garcinia myrtifolia (raubu), Firmania diversifolia
(cara), Geissois imthurnii (vo’a) and Dysoxylum sp. (tarawau kei soge). This composition is
somewhat different from rainforests on Viti Levu visited by me. On ridges Agathis macrophylla,
and Gymnostoma vitiense
may be common.
In the northernmost corner of Kubulau district and the reserve is an extensive plateau that has many of
the species found in the rainforest but is dominated by Atuna racemosa
and Fagraea gracilipes
) are also abundant. This plateau also has an abundance of
plants with stilt and prop roots, including Physokentia rosea (niuniu), Crossostylis spp. (no name
recorded), Myristica macrantha
) and Syzygium sp. (yasiyasi
). As flowing water and pools of
standing water are very common, this could also be considered a wetland.
Composition of Lowland Tropical Rainforest
A total of 839 trees belonging to 91 species in 43 families were encountered in the four plots totaling 1
ha. Their combined basal area (as calculated from the diameter at breast height (dbh)) was 45.6 m
Dominant species in terms of basal area were the flowering plants Myristica gillespieana
and Calophyllum vitiense
, and the conifer Retrophyllum vitiense
(Table 1). These four
species are mostly canopy species (M. gillespieana is also abundant in the subcaopy). However,
was the most common tree, reflecting its high abundance in the understorey
(Table 2). Other common species include Myristica gillespieana and Pagiantha thurstonii.
Table 1: The 30 most common species as measured by the basal area of individuals with a dbh of
10cm or more.
Table 2: The 30 most common species as measured by the number of individuals with a dbh of
10cm or more.
At the family level, the Myrisitcaceae (3 species of Myristica) and Clusiaceae (species of Calophyllum
) are dominant in terms of basal area (Table 3). Other important families in terms of basal
area are the Podocarpaceae (represented only by Retrophyllum vitiense), Anacardiaceae (Buchanania
, Pleiogynium timoriense
, Semecarpus vitiense
) and Chrysobalanaceae (almost entirely
eith a single individual of Atuna racemosa
). Gironniera celtidifolia
is so abundant
that the Ulmaceae is the most abundant family (Table 4). The Myrisitcaceae, Clusiaceae, Myrtaceae
(several species of Syzygium) and Apocynaceae (Pagiantha thurstonii and Alstonia costata) are other
Table 3: The 20 most common families as measured by the basal area of individuals with a dbh of
10cm or more.
10cm or more.
Table 4: The 20 most common families as measured by the number of individuals with a dbh of
A total of 319 species in 223 genera and 99 families were recorded (Table 5). All taxa recorded are
listed in appendix 1. This list is far from complete as several vegetation types (Coastal vegetation,
wetlands and anthropogenically induced vegetation types were not sampled). In addition, I expect
many additional taxa to be discovered in the lowland rain and mesic forests, once these have been more
thoroughly investigated. Of the 319 recorded species 288 were indigenous to Fiji. The low number of
introduced species is caused by the focus of my survey on native flora and vegetation. A total of 126
species were endemic to Fiji and 15 of those are restricted to Vanua Levu. The dicotyledons are the
biggest taxon, contributing more than two-thirds of all species and more than 90% of all endemic
The most important plant discovered was the small tree Zanthoxylum myrianthum (Rutaceae), which is
endemic to Vanua Levu and was recorded for the second time ever and for the first time in more than
50 years. In addition, the find of Astronidium kasiensis outside its only known and highly disturbed
location in the Mt. Kasi region is important. Other relatively narrow endemics include Veitchia filifera
(Arecaceae) Parsonsia smithii (Apocynaceae), Cyrtandra harveyi and C. reticulata (Gesneriaceae),
(Melastomataceae), Endospermum robbieanum
, Macaranga membranacea
Euphorbiaceae), Amaracarpus muscifer and Ixora coronata (both Rubiaceae), which are endemic to
Vanua Levu. Balaka seemannii and Physokentia thurstonii (both Arecaceae) are endemic to Vanua
Levu and Taveuni, while Cyathocalyx stenopetalus (Annonaceae) Cyrtandra dolichocarpa
(Gesneriaceae) are endemic to Vanua Levu and Rabi.
In addition, I identified 60 native taxa that were previously collected in the adjacent Mt. Kasi region
and Wainunu Catchment but were not recorded in this study. Of those 42 were endemic, 29 to Fiji, 7 to
Vanua Levu and six to either the Mt. Kasi region (Caesaria myrsinoides [Flacourtiaceae], Elaeocarpus
[Elaeocarpaceae], Mapania vitiensis
[Orchidaceae], Mapania vitiensis
[Myrtaceae], Phreatia flavovirens
[Orchidaceae]) or the Wainunu catchment
(Guioa capillacea [Sapindaceae]). A search whether the 6 narrowly endemic species are present in the
proposed reserve should be considered a priority as this would add great conservation value.
Four specimens collected are of special interest and potentially important records. While the
appearance of the specimens suggests them to be unique taxa, proper identification are required to
ascertain their identity. A specimen that I believe to belong to the genus Terminalia was collected in
fruit in one of the plots in the lowland rain forest. If my perception is correct, this specimen represents
a new species and the first non-coastal taxon with fleshy fruits. In addition, a specimen in the genus
was interesting as its leaves are covered with hairs, which is different from other species known
in Fiji. Using the key in Pannell (1992), the species comes closest to Aglaia tomentosa, which is known
only from New Guinea. An expert would need to determine whether the species is a first Fiji-record for
A. tomentosa, a new taxon, or just a form of species already described from Fiji. Another interesting
specimen was obtained in the genus Dolicholobium. Although the specimen was sterile, the indument
corresponds most closely to Dolicholobium aneityense, which is believed to be endemic to Vanuatu. It
will require fertile specimens and an expert in this genus to definitely identify the Dolicholobium.
Table 5: Floristic Summary of the Flora of Kubulau.
Numbers in the column refer to the number of
species. * = percentage endemism (endemic speces/ indigenous species × 100) is stated in paranthes
behind the number of species. ^ = includes endemic species.
Ferns & Fern
5 0 0
225 117(56.8) 12
Status of Forests
Before commenting on the status of the forests on the land of the mataqali Nadicake-Kilaka, I need to
point out that much of the present forest consists of patches of current plantations, plantations of the
recent past (grassland), plantations and village sites of the distant past (secondary forest), and intact
“old-growth” forest. A detailed survey of the entire land of the mataqali Nadicake-Kilaka should be
undertaken to identify those patches. In my opinion proper management of used, recovering and old-
growth areas is essential for the long-term protection of the remaining forests.
The status of the remaining and extensive forest patches is generally excellent. They include very large
individuals of slow-growing conifers (and faster growing flowering plants), attesting to their age (table
6). Not only is an individual of Agathis macrophylla with about 1.5m in diameter the biggest tree that I
have observed in Fiji, but also are many measurements in table 1 the largest recorded by me for the
particular species. The population of Retrophyllum vitiense is by far the biggest that I have ever
observed. In several forest fragments and on the plateau a high number (based on call frequency) of
tree frogs was recorded. As frogs are good indicator species for forest health (C. Morrison, pers.com.),
this also attests to the quality of the forest.
Table 6: Maximum trunk diameters (dbh) of species on the land of the mataqali Nadicake-
Kilaka, Kilaka, Kubulau, Bua.
damanu drau lailai
tarawau kei soqe
tabua mei ra kalavo Apocynaceae
No name recorded