Professional Documents
Culture Documents
1Swedish University of Agricultural Sciences, Faculty of Forestry, Department of Forest Vegetation Ecology,
S-901 83 Umeå, Sweden; Fax +46 90 166612; 2Permanent address: Alemaya University of Agriculture,
Faculty of Forestry, P. O. Box 138, Dire Dawa, Ethiopia; Tel. +251 5 111399
Abstract. The soil seed bank was investigated in four dry Somalia in the east, and from the Red Sea in the Sudan
Afromontane forests of Ethiopia. At least 167 plant species Republic in the north to the Cape Peninsula in the south
were identified in the 0 - 9 cm soil layer with total densities (White 1983). A large portion of the Afromontane area
ranging between 12 300 and 24 000 seeds/m2. Herbs were in Ethiopia was once covered by evergreen forests, but
represented with the largest numbers of species and seeds in
only fragments remain and even these are at constant
the seed bank, while the contribution of tree species was
generally low. The overall vertical distribution of seeds was risk of destruction. If the present trend of deforestation
similar at all sites with the highest densities occurring in the continues, the remaining forests will disappear in the
upper three cm of soil and gradually decreasing densities with very near future (Pohjonen & Pukkala 1990; Demel
increasing depth. Relatively high densities also occurred in the 1992; Tamrat 1993, 1994).
litter layer. There were large differences in depth distribution To manage the last remnants of Afromontane forests
between species, suggesting differences in seed longevity. A and to make a future restoration programme possible, it
large number of species in dry Afromontane forests evidently is essential to obtain information on the regeneration
store quantities of seeds in the soil and this is in contrast to the ecology of at least the dominant species of the vegeta-
situation in most tropical rain forests, dry lowland forests and
tion. Data on soil seed bank characteristics such as
savannas, where both the number of seeds and the number of
species are relatively small. It is possible that the strongly species number, seed quantity and depth distribution
seasonal and unpredictable climate of this region may have can provide insight into the regeneration ecology and
selected for high levels of dormancy, and that herb regenera- will in particular give an estimate of the regeneration
tion is associated with small scale disturbance. The fact that potential after disturbance (Thompson & Grime 1979;
most of the dominant tree species do not accumulate seeds in Grime 1989; Whitmore 1983, 1991; Fenner 1985;
the soil suggests that their regeneration from seed would be Garwood 1989; Thompson 1992). Studies in various
unlikely if mature individuals disappeared. Most tree species habitats have confirmed that, for species with a long-
have relatively large seeds and poor long-distance dispersal; lived seed bank, successful regeneration from buried
this implies that restoration of Afromontane forests after de-
seeds usually occurs after disturbance (Bakker 1989;
struction would be difficult. Since there is a diverse seed bank
of the ground flora, this component of the vegetation would Cavers & Benoit 1989; Keddy et al. 1989; van der Valk
have a better chance of re-establishing. However, because & Pederson 1989). If there is no long-lived seed bank,
most cleared forest land is used for agricultural crop produc- such regeneration takes place either through recently
tion, it is probable that the seed bank will be depleted in only a dispersed seeds, or, in the case of undisturbed vegeta-
few years. Therefore, the future of the Afromontane forest tion, through slow-growing seedlings and saplings.
flora seems to depend on the successful conservation of the Little is known about regeneration processes and
few fragments of remaining natural forest. soil seed banks in dry Afromontane forests, and infor-
mation on fruiting phenology, dispersal, longevity, pre-
dation and germination ecology is almost lacking, even
Keywords: Floristic composition; Germination; Regenera-
tion; Soil seed density; Soil seed distribution. for most of the dominant tree species. This lack of
knowledge of indigenous trees may have contributed to
the selection of Eucalyptus, Cupressus and other exotic
Nomenclature: Cufodontis (1953-1972); Hedberg & Edwards species for the few reforestation programmes which
(1989); Friis (1992). have so far been undertaken (von Breitenbach 1961a, b;
Pohjonen & Pukkala 1990; Demel 1993; Tamrat 1993).
The present study is part of a research project deal-
Introduction ing with various aspects of regeneration and seed ecol-
ogy in dry Afromontane forests. Here, we present an
The Afromontane Region extends intermittently at analysis of the soil seed bank at four widely separated
altitudes above 2000 m from Sierra Leone in the west to forest stands in the central and eastern highlands of
778 Demel Teketay & Granström, A.
Ethiopia, while concentrating on the question whether covers an area of ca. 500 ha. It was originally dominated
there is a persistent soil seed bank in dry Afromontane by Juniperus procera and Podocarpus falcatus, but
forests that could contribute to the restoration of natural most of the larger Juniperus trees have been removed
forest vegetation after severe disturbance. through selective cutting since the 1950s. The floristic
composition of the forest has been reported by Uhlig &
Uhlig (1990) and human impact has been described
Material and Methods elsewhere (Demel 1992).
March to April and from June to September, with very was inspected for seeds, which were identified to spe-
high precipitation in July and August. The size of the cies if possible, using local reference material. Seed
forest is difficult to ascertain; estimations range from viability was assessed by dissection; seeds were consid-
3500 - 10 000 ha (von Breitenbach 1962; Anon. 1991). ered viable if their content was white and firm. To
According to the inhabitants, there has not been any determine the habitat requirements of the species recov-
commercial exploitation of the forest. The phyto- ered from the soil seed bank, field observations were
sociology and structure of the forest have been de- made during trips in 1992 and 1993.
scribed by Tamrat (1993) while the floristic composi-
tion is put into relation with biodiversity issues by Data processing
Demel & Tamrat (1995).
Vertical distribution of seeds of a species is assumed
Soil sampling and analysis to reflect the longevity of its seeds in the soil (Leck
1989; Kjellsson 1992). To get a rough idea about the
At each site 10 plots were selected for soil sampling in ability of the species to survive in the soil, average depth
the closed part of each forest in order to roughly cover the distribution was calculated for each species using the
altitudinal range of each site. The sample plots at each site following formula:
were always at least 100 m apart. At each plot an area of 15
cm × 15 cm was marked and four separate soil layers were (SL × 0.5) + (ST × 2.5) + (SM × 5.5) + (SB × 8.5)
Average depth =
removed using a sharp knife and a spoon and put into cloth Total number of seeds
bags; these layers include litter layer and three succes-
sively deeper mineral soil layers, each 3 cm thick. Sam- where SL = number of seeds in the litter layer;
pling at all sites was done in the second half of March, at ST = number of seeds in the first soil layer (0 - 3 cm);
GA in 1991 and at the other three sites in 1992. Germina- SM = number of seeds in the second soil layer (3 - 6 cm);
tion trials were started within 1 - 2 weeks after sampling. SB = number of seeds in the third soil layer (6 - 9 cm).
The GA samples were incubated in a greenhouse in Umeå The multipliers 0.5, 2.5, 5.5 and 8.5 are the depth (in cm)
(Sweden) set at a day/night fluctuating temperature re- from the litter surface to the middle of the litter layer,
gime of 20/12 °C, which is a typical temperature range first, second and third soil layers, respectively. The litter
between day and night at the soil surface within the GA layer varied in depth but in these calculations an average
forest (Demel unpubl.). On sunny days in the summer, depth of 0.5 cm has been assigned.
temperature in the greenhouse reached up to 30 °C. The To compare similarities among sites, both in terms
samples received 12 h of light each day from Osram HQIT of species composition and number of seeds, the data
400 W fluorescent tubes except during the summer when were analysed using Principal Component Analysis with
there were long hours of daylight. The samples from the the computer program SIMCA-S (Anon. 1992-1994). In
other three sites were incubated in a lath-house at the the analysis, only taxa identified to species level and
Alemaya University of Agriculture, Ethiopia (altitude with at least 10 seeds recovered from the soil were
ca. 2000 m). Here, the samples were shielded from direct included. Species which were found to be outliers or
sunlight, but exposed to the ambient temperature which which did not contribute to the variation in the data were
ranged between 15 - 34 °C during the day and 8 - 15 °C excluded from the analysis. Sites were used as dummy
during the night. Soil samples were in all cases spread to variables in the analysis. Before the analysis, the number
a thickness of ca. 1 cm on cotton cloth in plastic trays and of seeds of each species in each sample was first con-
kept continuously moist. verted into a percentage value in relation to the highest
Seedlings started to emerge after about a week, and number of seeds of that species. The percentage values
those which were readily identifiable were recorded and were then transformed to the 1 - 9 scale similar to that of
discarded. Those difficult to identify at the seedling van der Maarel (1979) as follows: < 1 % seed (s) = 1; 2
stage were transplanted and grown separately until they % = 2; 3 % = 3; 4 % = 4; 5 - 12.5 % = 5; 12.5 - 25 % = 6;
could be identified. Sometimes this required flowering 25 - 50 % = 7; 50 - 75 % = 8 and > 75 % = 9.
of the plants. They were then pressed, and later identi-
fied at the Alemaya University of Agriculture Her-
barium or at the National Herbarium in Addis Abeba. Results
Every eight weeks, the soil samples were stirred to
stimulate seed germination. Species composition and density of seeds in the soil
After six months of incubation, the germination
trials were stopped and the samples were sieved using a The total number of species recorded, excluding the
mesh size of 0.5 mm. All material retained on this mesh unidentified ones, was 92 from GA, 66 from ME, 62
780 Demel Teketay & Granström, A.
from WW and 58 from MS (App. 1). The total number Table 2. Soil seed density (seeds/m2) of species common to all
of species from the four sites was 167, of which 119 the sites and the five species with the highest soil seed density
were dicotyledonous herbs, 20 were shrubs, 12 grasses (*) at each site.
and/or sedges, nine climbers and seven trees (Table 1). Species Munessa- Gara Ades Menagesha Wof-
11 species occurred at all four sites (Table 2), 21 at three Shashemene Washa
sites, 35 at two sites and 100 at one site (App. 1). The
Crassula alsinoides 1556* 4222* 782* 2076*
most well represented family was Asteraceae with 18 Cyperus rotundus - - 960* -
species at GA, 14 species at ME, 12 species at WW and Dichrocephala integrifolia 329 1089* 142 338
11 species at MS. Five species of ferns indigenous to Droguetia iners 1440* 218 - 489
Erica arborea - - 1142* 5293*
Ethiopia were also recorded. Geranium simensis 22 84 138 80
The number of viable seeds in the soil samples (from Haplocarpha rueppellii - - - 2587*
both germination trial and sieving) correspond to a seed Helichrysum foetidum 76 1116* 9 116
Lobelia giberroa 2236* 218 422 -
bank density down to 9 cm in the soil of 12300/m2 at Monopsis stellarioides 440 147 1649* 484
ME, 13700/m2 at MS, 20100/m2 at GA and 24000/m2 at Oxalis corniculata 467 31 391 400
WW. Quantitatively, the seed bank was dominated by Pilea tetraphylla 1560* 53 49 -
Plantago palmata 369 - - 1382*
relatively few species (App. 1). The five species with Poa annua 169 4 27 1409*
the highest seed density contributed 57 % of the total P. leptoclada 31 200 262 404
estimated seed bank at MS, 51 % at GA, 42 % at ME and Rubus apetalus 36 391 18 84
Solanum luteum 1022* 106 - -
53 % at WW (Table 2). Erica arborea, Haplocarpha S. nigrum 22 9 18 27
rueppellii and Crassula alsinoides had the highest aver- Veronica abyssinica 236 1929* 578* 9
age number of seeds (App. 1); for instance, E. arborea V. javanica 49 1867* 418 -
had 5293 and 1142 seeds/m2 of seeds at WW and ME,
respectively. Herbs had the highest number of species
and seeds compared with other life forms (Table 1). Vertical distribution of seeds
Seeds of woody species made up only a minor compo-
nent of the seed bank at all sites (except for the shrub The vertical distribution of the seed bank was similar
Erica arborea) and very few trees had considerable at the four sites, with the highest densities in the upper
amounts of viable seeds in the soil, i.e. Croton macro- 3 cm of soil and gradually decreasing densities with
stachyus (average = 33 seeds/m2 for sites where its seeds increasing depth; relatively high densities also occurred
were recovered; App. 1), Juniperus procera (127 seeds/ in the thin litter layer (Fig. 1a). The number of species
m2), Nuxia congesta (271 seeds/m2) and Ficus sur (498 present showed a similar but less pronounced trend (Fig.
seeds/m2). 1b). There was more variation between species than
In the PCA ordination, the sites were clearly sepa- between sites in relation to the vertical distribution. For
rated along the first and third principal component axes. instance, seeds of some species were almost entirely
The first component separated GA and WW while the confined to the litter layer, e.g. Juniperus procera,
third component separated MS and ME. The first and Clematis hirsuta, Girardinia diversifolia and Pilea
second components, could not separate MS and ME, tetraphylla. On the other hand some species have seeds
indicating that the two sites are very similar to each which are well distributed in all soil layers, e.g. Laggera
other and more so than they are to GA and WW in their crispata, Lobelia giberroa, Crassula alsinoides, Ve-
species composition and quantity of seeds in the soil. ronica abyssinica and Poa leptoclada, while some oth-
ers had most seeds distributed in the deeper layers, e.g.
Indigofera rothii, Solanum nigrum and Eragrostis
schweinfurthii. The average depth of seeds of the differ-
Table 1. Number of species in the different life forms at the ent species gives an indication of their distribution in the
four sites. soil (Fig. 2, App. 1). There were some differences re-
Life form Number of species garding life forms, with herbs, grasses and sedges found
Munessa- Gara Ades Menagesha Wof-Washa
at greater depths than trees, shrubs and climbers (Fig. 2).
Shashemene As mentioned earlier, woody plants generally had
Trees 2 7 2 -
low seed numbers in the soil, but there were some
Shrubs 6 12 8 3 notable exceptions. Even though Ficus sur is repre-
Climbers 4 6 3 2 sented by only very few adult trees in the forest at GA, it
Herbs 44 59 47 53
Grasses and sedges 2 8 6 4
had the highest number of seeds (498 seeds/m2) among
the trees, and these seeds were well distributed in the
Total 58 92 66 62 deeper layers. Most seedlings of this species emerged
- Soil seed banks in dry Afromontane forests of Ethiopia - 781
Fig. 1. Vertical distribution of seeds (a) and species (b) at the four sites (depth refers to the midpoints of the sampled layers).
Table 3. Apparently viable (VI), eaten (EA) and dead (DE) seeds sieved from soil samples after the germination experiments
terminated.
Trees
Croton macrostachyus 1 - 1 2 - 2 - 1 - - - -
Juniperus procera 3 - 90 45 141 534 17 42 549 - 26 40
Olea europaea subsp. cuspidata - - - 4 1 20 - 2 10 - - -
Podocarpus falcatus - 3 43 2 - 4 - - - - - -
Rhus sp. - - - - - 2 - - - - - -
Shrubs
Indigofera rothii - - - 14 - - - - - - - -
Rubus sp. - - 49 4 - 26 - - - - - 20
Climbers
Clematis sp. - - - - - 1 - - - - - -
Galium sp. - - - 3 - 6 - - - - - -
Herbs
Chenopodium sp. - - - 4 - 5 - - - - - -
Geranium sp. - - - 1 - - - - - - - -
Girardinia diversifolia 74 - - - - - - - - - - -
Trifolium sp. - - - - - - - - - 38 - -
Unidentified
Sp. A - - - - - - 1 1 2 - - -
Sp. B - - - - - - - - - 2 - -
Sp. C - - - - - - - - - 7 - -
Sp. D - - - - - - - 3 11 - - -
Others - - - 7 18 62 - - - - - -
excluded. A low-density seed population in the soil does Hagenia abyssinica, Bersama abyssinica and Ekebergia
not by itself infer short persistence, but the limit of five capensis, failed to reveal the presence of viable seeds.
seeds was adopted as a reasonable sample size for the This general lack of a soil seed bank in tree species may
depth distribution. We have also attempted to group be due to several factors. Some species, notably B.
species into four categories with respect to their habit, abyssinica and E. capensis, appear to have recalcitrant
based on observations in the field and literature, i.e. seeds, adapted to germination more or less immediately
shade-intolerant/gap-colonizing species (G), indifferent after seed shed (Demel unpubl.). Others may have a
(establishing both in gaps and shade) (I), shade-tolerant capacity to store seeds in the soil for several years. In
(found mainly under shade of trees and shrubs) (S), and burial experiments 99 % and 93 % of the seed popu-
uncertain (no information about their preferences) (U). lations of Acacia abyssinica and Croton macrostachyus,
The majority of the species with persistent seed banks respectively, were still alive after 37 months of burial in
(58 %) are gap colonizers (App. 1). Several species, e.g. the soil (Demel unpubl.). It is likely that predation of
Poa annua, Solanum nigrum, and Stellaria media, are these large seeds on the soil surface prohibits the build-
weeds and ruderal species of Eurasian origin which have up of a substantial seed bank. Pudden (1958) reported
become cosmopolitan. from Kenya that large quantities of seed of Juniperus
The data show that the majority of woody plant procera were eaten by monkeys in the tree canopies,
species do not accumulate seeds in the soil. Previous and about 75 % of the seeds reaching the ground were
floristic studies of the standing vegetation at the four consumed by rodents.
sites show that there are at least 66 species of woody Although we have no quantitative data on seed pre-
plants, including woody climbers, at the Munessa- dation, inspection of rodent nests at GA revealed large
Shashemene forest (Lundgren & Lundgren 1969), 58 at heaps of seed coats of both Olea europaea ssp. cuspidata
Gara Ades forest (Uhlig & Uhlig 1990; pers. observ.), and Podocarpus falcatus. A considerable number of
55 at Menagesha forest (Sebsebe 1988) and 54 at Wof- eaten seeds of J. procera and a small number of seeds of
Washa forest (Demel & Tamrat 1995). In several in- O. europaea ssp. cuspidata and P. falcatus (Table 3)
stances, soil samples taken virtually under the canopy of were also found in the soil which was sieved after the
large mature trees of many common species, e.g. germination trial was terminated, suggesting that preda-
Podocarpus falcatus, Olea europaea ssp. cuspidata, tion may play an important role in reducing the seed
- Soil seed banks in dry Afromontane forests of Ethiopia - 783
montane forest flora seems to depend on the successful of Wof-Washa natural forest, Central Ethiopia: implica-
conservation of the few remaining fragments of natural tions for the conservation of biodiversity. Feddes Repert.
forest. 106: 127-147.
Enright, N. 1985. Existence of a soil seed bank under rain
forest in New Guinea. Aust. J. Ecol. 10: 67-71.
Acknowledgements. Financial support for this project was Fenner, M. 1985. Seed Ecology. Chapman and Hall, London.
obtained from the Swedish International Development Agency Friis, I. 1992. Forests and Forest Trees of Northeast Tropical
and the Alemaya University of Agriculture. Logistic support Africa. Kew Bull. Add. Ser. 15: 1-396.
was provided by the Faculty of Forestry, Faculty of Agricul- Garwood, N.C. 1989. Tropical Soil Seed Banks: A Review.
ture and Debre-Zeit Research center (AUA), State Forest In: Leck, M.A., Parker, V.T. & Simpson, R. L. (eds.)
Offices at the four sites, ACD and ETH. We thank Olle Ecology of Soil Seed Banks, pp. 149-209. Academic Press,
Zackrisson, David Wardle and three anonymous reviewers for San Diego, CA.
comments on the manuscript, Anders Nordgren and Christer Gilbert, E.F. 1970. Mount Wachacha: A botanical commen-
Albano for their help in the multivariate analysis, Sylvia tary. Walia 2: 3-12.
Phillips for identifying grasses, Ove Ericsson for assisting in Graham, A.W. & Hopkins, M.S. 1990. Soil seed banks of
data handling, Tamerat Abebe, Abdurazak Abdulahi, Mekdes adjacent unlogged rain forest types in north Queensland.
Mulu, Dawit Mulu and Keflie Lebassie for their help in the Aust. J. Bot. 38: 261-268.
field and lath-house. Grime, J.P. 1989. Seed Banks in Ecological Perspective. In:
Leck, M.A., Parker, V.T. & Simpson, R.L. (eds.) Ecology
of Soil Seed Banks, pp. xv-xxii. Academic Press, San
References Diego, CA.
Guevara, S.S. & Gomez-Pompa, A. 1972. Seeds from surface
Anon. 1988. National Atlas of Ethiopia. Ethiopian Mapping soils in a tropical region of Veracruz, Mexico. J. Arnold
Authority, Addis Abeba. Arbor. 53: 312-335.
Anon. 1991. Biodiversity Guide to Ethiopia. World Conserva- Hall, J.B. & Swaine, M.D. 1980. Seed stocks in Ghanaian
tion Monitoring Centre, Cambridge. forest soils. Biotropica 12: 256-263.
Anon. 1992-1994. User’s Guide to SIMCA-S. Umetri AB, Hedberg, I. & Edwards, S. 1989. Flora of Ethiopia, Vol. 3.
Umeå. The National Herbarium, Addis Abeba.
Alvarez-Buylla, E.R. & Martinez-Ramos, M. 1990. Seed bank Hopkins, M.S. & Graham, A.W. 1983. The species composi-
versus seed rain in the regeneration of a tropical pioneer tion of soil seed banks beneath lowland tropical rain
tree. Oecologia (Berl.) 84: 314-325. forests in north Queensland. Biotropica 15: 90-99.
Bakker, J.P. 1989. Nature Management by Grazing and Cut- Keay, R.W. 1960. Seeds in forest soils. Niger. For. Inf. Bull. 4:
ting. Kluwer Academic Publishers, Dordrecht. 1-4.
Cavers, P.B. & Benoit, D.L. 1989. Seed banks in Arable Land. Keddy, P.A., Wisheu, I.C., Shipley, B. & Gaudet, C. 1989.
In: Leck, M.A., Parker, V.T. & Simpson, R.L. (eds.) Seed Banks and Vegetation Management for Conserva-
Ecology of Soil Seed Banks, pp. 309-328. Academic Press, tion: Toward Predictive Community Ecology. In: Leck,
San Diego, CA. M.A., Parker, V.T. & Simpson, R.L. (eds.) Ecology of Soil
Chaffey, D.R. 1979. South-west Ethiopia forest inventory Seed Banks, pp. 347-363. Academic Press, San Diego,
project: a reconnaissance inventory of forest in south- CA.
west Ethiopia. Ministry of Overseas Development, Land Kjellsson, G. 1992. Seed banks in Danish deciduous forests:
Resource Development Centre, Project Report 31, Lon- species composition, seed influx and distribution pattern
don. in soil. Ecography 15: 86-100.
Chaffey, D.R. 1980. South-west Ethiopia forest inventory Leck, M.A. 1989. Wetland seed banks. In: Leck, M.A., Parker,
project: an inventory of forest at Munessa and Shashemene. V.T. & Simpson, R.L. (eds.) Ecology of Soil Seed Banks,
Ministry of Overseas Development, Land Resource Divi- pp. 283-305. Academic Press, San Diego, CA.
sion, Project Report 29, London. Liew, T.C. 1973. Occurrence of seeds in virgin forest top soil
Cufodontis, G. 1952-1973. Enumeration Plantarum Aethiopiae with particular reference to secondary species in Sabah.
Spermatophyta. Bull. Jard. Bot. Etat Brux. Vols. 23-42. Malay. For. 36: 185-193.
Bruxelles. Lundgren, B. 1971. Soil studies in a montane forest in Ethio-
Daniel Gamachu 1977. Aspects of Climate and Water Budget pia. Royal College of Forestry, Department of Forest
in Ethiopia. Addis Abeba University Press, Addis Abeba. Ecology and Forest Soils, Research Notes No. 11, Stock-
Demel Teketay 1992. Human impact on a natural montane holm.
forest in south-eastern Ethiopia. Mount. Res. Devel. 12: Lundgren, L. & Lundgren, B. 1969. The Munessa Forest: A
393-400. plant ecological study. Minor Research Tasks 2, CADU,
Demel Teketay 1993. Problems associated with raising trees Addis Abeba.
from seeds, the Ethiopian experience. In: Lieth, H. & Mohr, P.A. 1971. The Geology of Ethiopia (2nd ed.). Univer-
Lohmann, M. (eds.) Restoration of Tropical Forest Eco- sity College of Addis Abeba Press, Addis Abeba.
systems, pp. 91-100. Kluwer, Dordrecht. Murphy, H.F. 1968. A report on the fertility status and other
Demel Teketay & Tamrat Bekele 1995. Floristic composition data on some soils of Ethiopia. College of Agriculture,
- Soil seed banks in dry Afromontane forests of Ethiopia - 785
HSIU Experimental Station Bull. No. 44, Alemaya. Negro region of the Amazon Basin. J. Ecol. 69: 631-649.
Pohjonen, V. & Pukkala, T. 1990. Eucalyptus globulus in Uhlig, S.K. & Uhlig, K. 1990. The floristic composition of a
Ethiopian forestry. For. Ecol. Manage. 36: 19-31. natural montane forest in south-eastern Ethiopia. Feddes
Pudden, H.H.C. 1958. Natural regeneration (of Juniperus Repert. 101: 227-234.
procera). Rep. For. Dep. Kenya 1955/1957: 15-16. van der Maarel, E. 1979. Transformation of cover-abundance
Putz, F. E. & Appanah, B. 1987. Buried seeds, newly dis- values in phytosociology and its effects on community
persed seeds, and the dynamics of a lowland forest in similarity. Vegetatio 39: 97-114.
Malaysia. Biotropica 19: 326-339. van der Valk, A.G. & Pederson, R.L. 1989. Seed Banks and
Raunkiaer, C. 1934. The Life-forms of Plants and Statistical the Management and Restoration of Natural Vegetation.
Plant Geography. Clarendon Press, Oxford. In: Leck, M.A., Parker, V.T. & Simpson, R.L. (eds.)
Russell-Smith, J. & Leucas, D.E. 1994. Regeneration of Ecology of Soil Seed Banks, pp. 329-346. Academic Press,
monsoon rain forest in northern Australia: the dormant San Diego, CA.
seed bank. J. Veg. Sci. 5: 161-168. von Breitenbach, F. 1961a. Forests and woodlands of Ethio-
Saulei, S.M. & Swaine, M.D. 1988. Dynamics of seed rain and pia, a geobotanical contribution to the knowledge of the
soil seed banks during forest clearance and subsequent principal plant communities of Ethiopia, with special re-
regrowth in the Gogol Valley, Papua New Guinea. J. Ecol. gard to forestry. Ethiop. For. Rev. 1: 5-16.
76: 1133-1152. von Breitenbach, F. 1961b. Exotic Trees in Ethiopia. Ethiop.
Sebsebe Demissew 1988. The floristic composition of For. Rev. 2: 19-39.
Menagesha State Forest and the need to conserve such von Breitenbach, F. 1962. National Forestry Development
forests in Ethiopia. Mount. Res. Devel. 8: 243-247. Planning, a feasibility and priority study on the examples
Skoglund, J. 1992. The role of seed banks in vegetation of Ethiopia. Ethiopian For. Rev. 3/4: 41-68.
dynamics and restoration of dry tropical ecosystems. J. von Breitenbach, F. & Koukol, J. 1962. Menagasha State
Veg. Sci. 3: 357-360. Forest, a description of the forest, its management and its
Tamrat Bekele 1993. Vegetation ecology of remnant Afro- future development including the National Park Project.
montane forests on the Central Plateau of Shewa, Ethio- Ethiop. For. Rev. 3/4: 17-34.
pia. Acta Phytogeogr. Sue. 79: 1-59. White, F. 1983. The Vegetation of Africa. A Descriptive Memoir
Tamrat Bekele 1994. Phytosociology and ecology of a humid to Accompany the UNESCO/AETFAT/UNSO Vegetation
Afromontane forest on the Central Plateau of Ethiopia. J. Map of Africa. UNESCO, Paris.
Veg. Sci. 5: 87-98. Whitmore, T.C. 1983. Secondary succession from seed in
Thompson, K. 1987. Seeds and seed banks. New Phytol. 106 tropical rain forests. For. Abstr. 44: 767-779.
(Suppl.): 23-34. Whitmore, T.C. 1991. Tropical Rain Forest Dynamics and Its
Thompson, K. 1992. Functional ecology of soil seed banks. In: Implications for Management. In: Gomez-Pompa, A.,
Fenner, M. (ed.) Seeds: The Ecology of Regeneration in Whitmore, T.C. & Hadley, M. (eds.) Rain Forest Regen-
Plant Communities, pp. 231-258. eration and Management, pp. 67-89. UNESCO, Paris.
Thompson, K. & Grime, J.P. 1979. Seasonal variation in the Young, K.R. 1985. Deeply buried seeds in a tropical wet forest
seed bank of herbaceous species in ten contrasting habi- in Costa Rica. Biotropica 17: 386-338.
tats. J. Ecol. 67: 893-921. Young, K.R., Ewel, J.J. & Brown, B.J. 1987. Seed dynamics
Uhl, C., Clark, K., Clark, H. & Murphy, P. 1981. Early plant during forest succession in Costa Rica. Vegetatio 71: 157-
succession after cutting and burning in the upper Rio 173.
App. 1. Quantity of viable seeds in soil samples from the four sites (germination trials and soil sieving combined); species identified
at the species or genus level and having more than five seeds are included. S = seed bank type; P = persistent seed bank; H = habit;
G = gap-colonizing; I = indifferent; S = shade-tolerant; U = unknown. AD = average depth of occurrence in cm.
S H AD MS GA ME WW S H AD MS GA ME WW