The vascular flora of 1.5 ha of montane rain forest on a ridge top at
2550-2650 m near Merida, Venezuela was surveyed, compared with other m
ontane forests in the region, and analysed to determine its biogeograp
hic origins and affinities. The study included an inventory of all ste
ms greater-than-or-equal-to 3.2 cm d.b.h. in eighteen 12 x 12 m plots.
The canopy was 10-22 m tall and the trees relatively small (3488 indi
viduals ha-1, basal area 34.7 m2 ha-1). Bamboos were abundant in gaps,
there were no buttressed trees and relatively few climbers. Weinmanni
a glabra L.f. (Cunoniaceae), Clusia trochiformis Vesque (Guttiferae) a
nd Hedyosmum creantum Occhioni (Chloranthaceae) together accounted for
42% of the individuals and 37% of the basal area. Microphylls, notoph
ylls and mesophylls predominated in the leaf size spectrum. Overall, 2
19 vascular plant taxa were distinguished in the 1.5 ha study area, in
cluding forty-four tree species (plus c. eight Lauraceae), fifteen shr
ubs, twenty-six herbs (including terrestrial ferns), twenty-one climbe
rs, eight hemiepiphytes, 120 true epiphytes and four hemiparasitic epi
phytes (many species occurred as more than one life form). Orchidaceae
(sixty-five species), Bromeliaceae (fourteen) and Rubiaceae (eleven)
were the best-represented families. The proportion of true epiphytes i
n the total flora (50%) exceeds estimates published for other sites. F
orty-four per cent of the flora of the study area is restricted to nor
thern Venezuela and Colombia, 25% is endemic to Venezuela and 7% is en
demic to Merida state. The flora is thus more geographically specializ
ed than those of isolated montane forests at lower altitudes near the
Caribbean coast. The majority of the flora (65%, principally epiphytes
) had wind-dispersed diaspores (dust, comose, winged) and 29% of the s
pecies (including most trees and shrubs) had succulent fruits. The spe
cies with succulent fruits had more limited distributions than the win
d-dispersed species. The forest of the study site is floristically dis
tinct from even the nearest other studied montane forests. Future stud
ies should examine whether inter-site differences in montane forests a
re greater than those between lowland sites separated by similar dista
nces.