Browse in northern Africa**

H. N. Le Houérou

Introduction

2. The main types of grazing lands and the importance of browse in animal feeding

2.1 General

2.2 Browse ecosystems

3. Introduced species

4. Forage value

4.1 Palatability

4.2 Voluntary intake

4.3 Chemical composition and nutritive value

4.4 Apparent digestibility coefficient of organic matter in some North African browse species (average values)

5. Productivity

5.1 General remarks

5.2 Production of browse in the semi-arid and humid bioclimatic zones

5.3 In the arid zone

6. Management

7. Conclusions

References

Introduction

Northern Africa, or Mediterranean Africa, as understood here, comprises the African countries lying north of the Tropic of Cancer i.e. Morocco, Algeria, Tunisia, Libya and Egypt. Unlike the other African countries, these are submitted to the Mediterranean climate characterized by autumn, winter and spring rains and summer drought, i.e. a rainfall regime opposite to that of the tropics, further south.

As a consequence of this, North African vegetation is of a Mediterranean type, totally different in its botanical composition, structure and physiognomy from any tropical African vegetation (albeit it bears some structural and physiognomical similarities with the afro-tropical montane vegetation of evergreens and also some kinship with the residual "Rand Flora" of the higher elevations).

Livestock are also different in species, races and numerical proportions in their composition. Cattle belong, in their large majority, to the Bos taurus species often more or less mixed with European stock and some recently imported Bos indicus stock from India and Pakistan (Nellore, Red Sindi and Sahiwal breeds). In relative numbers, cattle are much less important (14%) than in the tropics. Small stock represent 85% of total livestock numbers and camels less than 1%. Sheep are of a wool-bearing type, of the fat tail group to the east (Tunisia, Libya, Egypt), and of the narrow tail group to the west. Goats belong to the black so-called Nubian breed, that foresters sometimes label the "black locust".

2. The main types of grazing lands and the importance of browse in animal feeding

2.1 General

Livestock are, to a large extent, raised in a traditional way and fed on year-round grazing, with the exception of some dairy cattle in the irrigated areas and around the main urban centres, and also a few feedlots.

Grazing lands include rangelands, stubble, fallow, some very limited natural grassland and some sown pastures. Rangeland and cropland (stubble and fallow) probably account for 60 and 40% of livestock feeding respectively although the area of rangeland is three times larger than cropland.

Browse is a very important component of rangelands and accounts for at least 60% of its production, as an interannual average. It can thus be estimated that browse represents a minimum of 35 to 40 percent of livestock feed in Mediterranean Africa, as opposed to less than 20% in tropical Africa. This situation is due to a number of causes:

1. There are very few, and limited in extent, grass dominated natural ecosystems in the area (except for the esparto steppe of Stipa tenacissima, which is almost unpalatable).
2. Palatable perennial grasses and legumes are scarce in grazed natural ecosystems, having been eliminated by centuries of overstocking.
3. Grazed natural ecosystems are dominated by shrubs, both in the semiarid to humid ecological zones (rainfall >400 mm) where grazing lands are mostly degraded forests of the maquis or garngue type dominated by shrubs 0.5 to 5.0 m high, are in the arid (400 > R > 100 mm) and desert (100 mm > R) ecological zones where steppic ecosystems are usually dominated by chamaephytic undershrubs less than 0.5 m high.
4. 4) The bulk of animal feed from rangelands is approximately constituted by 30–40% of annual grasses, legumes and forbs, 55–65% of shrubs and trees and perhaps an optimistic 5% of perennial grasses and legumes. Of course, these proportions may vary largely from one type of rangeland to the next and from year to year in a given range type, according to rainfall amount and distribution. Annual species are more important in good rainfall seasons whereas browse is essential in time of drought. Therefore the role of browse increases with the climatic aridity, or, to put it the other way around: other conditions being equal, the role of browse is inversely proportional to the amount of rainfall, both on a seasonal or annual basis. The same thing is true for the tropics.

2.2 Browse ecosystems

As mentioned above, the large majority of North African rangelands are dominated by shrubs; most of those are browsed. These "browselands" cover the following surfaces in 10³ km²:

Table 1.North African browselands in 10³ km²

Country	Total	Semi-arid to humid rangelandszone (R > 400 mm) shrublands	Arid rangelands (400 > R > 100)chamaephytic steppe	Desert rangelands (100 > R > 50) chamaephytic steppe
Algeria	345.6	25.6	120.0	200.0
Egypt	125.0	–	25.0	100.0
Libya	240.7	0.7	80.0	160.0
Moroccoa	142.0	32.0	80.0	30.0
Tunisia	87.2	7.2	45.0	35.0
Total	940.5	65.5	350.0	525.0

^aAs from 1970.

There are, in addition to rangelands, some 333 000 km²(33 million hectares) of grazed stubble, fallows, fodder crops, meadows and sown pastures as shown in Table 2.

Table 2. Other grazing resources in 10³ km²

There is also an unknown acreage of planted fodder shrubs and trees (Cactus Acacia, Atriplex mainly), probably in the region of half a million hectares, 80% of which is spineless cactus. This is probably the largest expanse of artificially established browse in the world to-day.

The main browse ecosystems in North Africa and their approximate acreage are shown in Tables 3 and 4.

Table 3.The main browse ecosystems in the semi-arid to humid bioclimatic zones of northern Africa, in 10³ km²

Table 4. The main browse ecosystems in the arid bioclimatic zone of northern Africa and their estimated acreage

Table 5. Livestock populations in 10³ as from 1976

2.2.1 The semi-arid to humid bioclimatic zones

In the semi-arid to humid ecological zones, almost all forest lands (with the exception of very few and limited timber forests) are actual grazing lands. Most of those are degraded forests of the maquis or garrigue type (matorral or chaparral) where the main resource is grazing; secondary resources being fuel gathering, charcoal and distillation (rosemary, juniper). Even in the timber or cork-producing forest, grazing represents 60 to 80% of the revenue of the land (Le Houérou, 1971, 1972). The only exceptions are industrial plantations of eucalyptus, which amount to a few thousand hectares only.

In their immense majority, these shrublands are essentially browselands on which a livestock population of some 20 million sheep equivalent depend entirely for their feeding (out of a total livestock population of some 100 million sheep equivalents in the region); the remainder of the livestock population depending partly on rangeland and partly on cropland

2.2.1.1 The Cedrus atlantica series

This includes a series of plant communities derived from the Cedrus atlantica forest climax occurring on various substrata from granite to limestone. This series occurs in high mountains above 1200–1500 m and up to 2500–2800 m, with average rainfall above 600 mm; one to three months of snow cover and hard frost for 60 to 120 days; summer drought, however, lasts 2 to 4 months. Dominant or characteristic shrub and tree species are:

Cedrus atlantica, Taxus baccata, Juniperus communis, Fraxinus xanthoxyloides^b, Quercus ilex s.l.^b, Medicago suffruticosa^b, Acer campestre^a, Acer opaluse^a, Cotoneaster nummularia^a, Buxus balearica, Berberis hispanica^c, Bupleurum spinosum, Juniperus thurifera, Juniperus oxycedrus^a, Ilexaquifolium, Cytisus battandieri^b, Ulmus campestrirs^b, Acer monspessulanum^a, Acer obtusatum^a, Prunus prostrata^a, Amelanchier ovalis^a, Daphne laureola, Cytisus purgans, Erinacea anthyllis.

^a indicates browsed species, ^b indicates heavily browsed species, ^c indicates occasionally browsed species (often by goats and/or, camels only).

The Cedrus forest or shrubland is good summer range for some of the transhumant herds of Morocco and Algeria where they occur on some 150 000 hectares.

2.2.1.2 The Quercus faginea/ Q. suber series

This series occurs in high rainfall areas from 600 mm up to over 2000 mm. The Q. suber forests, matorral or maquis are often depleted stages of the Q. faginea climax (Quezel, 1954; Le Houérou, 1973). This series is linked to non calcareous substrata and develops on more or less acidic soils. The series covers some 1.3 million hectares in Algeria, Morocco and Tunisia. Dominant or characteristic shrub and tree species are:

Q.faginea^b, Cytisus triflorus^b, Olea europaef. oleaster^b, Celtis australis^b, Rhammus alaternus^b, Pistacia lentisus^a, Erica arborea^c, Calycotomevillosa^c, Ruscus hypophyllum, Tamus communis, Q. suber^b, Viburnum tinus^b, Phillyrea media latifolia^b, Fraxinus oxyphylla^b, Fraximus augustifolia^b, Arbutus unedo^a, Myrtus communis^c, Daphne gnidium, Smilax aspera^a.

Some species such as Fraxinus oxyphylla are pollarded and lopped branches and twigs are fed to livestock (in Kabylia of Algeria, and in the Rif of Morocco). Acorns of Q. faginea and Q. suber are searched for by all livestock and wild boars.

^a indicates browsed species, ^b indicates heavily browsed species, ^c indicates occasionally browsed species (often by goats and/or camels only).

2.2.1.3 The Quercus ilex s.l. series

This series (including here, for convenience, the Q. coccifera and Q. calliprinos series of minor importance) very rarely constitute fully grown forest or even parkland, but generally matorral shrubland in the semi-arid and subhumid Mediterranean bioclimates with average rainfalls ranging from 400 to 800 mm. They occur on various types of substrata and soils. They cover some 2.2 million hectares mainly in Algeria and Morocco. Dominant shrub species are:

Quercus ilex s.l. (including Q. rotundifolia)^b, Phillyrea angustifolia subsp. media^b, Rhamnus alaternus^b, Argyrolobiun linneanum^b, Coronilla emeroides^b, Coronilla minima^b, Ulmus campestris subsp. procera^b, Crataegus monogyna^b, Arbutus unedo^a, Pistacia terebinthus^a, Dorycnium suffruticosum^a, Genista spp, Erica scoparid^c, Rosmarinus officinalis^c, Rosmarinus tournefortii, Teucrium spp^c, Cistus spp., Rhamnus lycioides subsp. oleoides^b, Colutea arborescens^b, Coronilla juncea^b, Medicago suffruticosa^b, Crataegus laciniata^b, Juniperus oxycedrus^a, Pistacia lentiscus^a, Pistacia atlanttca^a, Spartium junceum^c, Erica arborea^c, Erica multiflora^c , Thymus sp.p, Satureja sp.p Asparagus acutifolius, Rubia peregrina.

This type of browseland covers some 21000 km² of which 13.5 are in Morocco 7.0 in Algeria 0.4 in Tunisia and 0.1 in Libya.

2.2.1.4 The Olive-mastic and Olive-carob series

These two series of vegetation usually occupy plains and lowland with deep soils and a favourable water balance in the semi-arid to humid bioclimates with mild winters (350 to 1000 mm annual rainfall). As a consequence, they have, to a very large extent, been cleared for farming during the course of history. For the past 22 centuries about 15 million ha (150 000 km²) where cleared for cultivation (Le Houérou, 1980). Most of those were probably covered with the Oleo-ceratonion and Oleo-lentiscetum to use the Zuricho-Montpellieran jargon. At the present time the Oleo-ceratonion and Oleo-lentiscetum have become almost vestigial, covering less than 7 200 km² of which 5 000 are in Morocco, 1000 in Algeria, 700 in Tunisia and 50 in Libya.

Major shrub species are:

Olea europeaf. oleaster^b, Phillyrea media^b, Rhamnus alaternus subsp. myrtifolia^b, Rhamnus lycioides subsp. oleoides^b, Periploca laevigata^b, Fraxinus oxyphylla^b, Cytisus mollis^b, Colutea arborescens^b, Prasium majus^b, Pistacia lentiscus^a, Prunus amygdalus^a, Prunus mahaleb^a, Rhus tripartita^c, Ziziphus lotus^c, Jasminum fruticans^c, Corydothymus capitatus, Genista spp. Anagyris foetida, Psoralea bituminosa, Clematis flammula, Ceratonia siliqua^b, Quercus coccifera^b, Coronilla emeroides^b, Fraxinus angustifolia^b, Crataegus oxyacantha^b, Lonicera implexa^b, Lonicera etrusca^b, Crataegus ararolus^a, Prunus avium^a, Helianthemum spp^a, Rhus pentaphylla^c, Calycotome villosa^c, Lycium europaeum^c , Cistus spp., Anthyllis barbs-jovis, Asparagus albus, Ulex spp. Nerim aleander.

2.2.1.5 The Aleppo pine series

The Aleppo pine series occupies huge areas in North Africa especially on calcareous stony and many hills of the semi-arid and subhumid bioclimates having cool to cold winters: 13 000 km², of which 8 500 are in Algeria, 3 500 in Tunisia, 700 in Morocco and 50 in Libya. The main shrub species are similar to those listed for the Quercus ilex series with however qualitative differences in the botanical composition and also quantitative differences in the proportion of the dominating species.

Important shrubs and undershrubs are:

Pinus halepensis, Olea europaea^b, Globularia alypum^b, Rhamnus lycioides subsp. oleoides^b, Colutea arborescens^b, Onobrychis argentea^b, Hippocrepis scabra^b, Fumena ericoides^a, Fumena thymifolia^a, Fumena laevipes^a, Thymelaea nitida^a, Juniperas phoenicea^c, Rosmarinus officinalis^c, Rosmarinus tournefortii^c, Rhamnus lycioides subsp. lycioides^c, Rhus tripartitum^c, Sparium junceum, Ruta chalepensis, Ruta montana, Cistus albidus, Teucrium spp. Retama sphaerocarpa, Atractylis humilis, Anarrhinum fruticosum, Dorychium suffruticosum, Quercus ilex^b, Phillyrea media^b, Coronilla minima^b, Coronilla juncea^b, Hedysarum naudinianum^b, Hedysarum humile^b, Hedysarum perralderianum^b, Helianthemum spp^a, Anthyllis sericea subsp. sericeae^a, Lotophyllus argenteus^a, Juniperus oxycedrus^c, Pistacia lentiscus^c, Cistus libanotis^c, Cistus salviaefolius^c, Genista cinerea, Genista microcephala, Genista tricuspidata, Genista quadriflora, Cistus crispus, Nupleurum gibraltaricum, Bupleurum balancae, Thymus hirtus, Linum umbellatum, Pituranthos scoparius.

2.2.1.6 The false thuya series

Occupies some 9 400 km², of which 7 400 are in Morocco, 1 600 in Algeria and 300 in Tunisia, in the semi-arid bioclimate with mild to warm winters (rainfall 350 to 600 mm, mean minimum of January above 3°C). Otherwise it occurs in the same geomorphological conditions as the Aleppo pine series i.e. on shallow soils of calcareous and many hills and low mountains. The main species are similar to the Aleppo pine series but with a few differential species both qualitatively and quantitatively. This series is closely akin to the Oleo-ceratonion in its botanical composition.

The main shrub species are:

Tetraclinis articulata, Olea europaea, f. oleaster^b, Periploca loevigata^b, Withania frutescensb, Globularia alypum^b, Teucrium fruticans^b, Artemisia herba alba, Phagnalon saxatile^a, Phagnalon rupestre^a, Rhus pentaphylla^c, Ziziphus lotus^a, Rosmarinus officinalis^c, Sarcopoterium spinosum^c, Calycotomevillosa^a, Calycotoma intennedia^c, Cistus libanotis^c, Cistus salviifolius^c, Cistus crispus^c, Asparagus albus, Genita cinerea, Sideritis romana, Convolvulus cantabrica, Arganica sideroxylon (Morocco only)^b, Ceratonia siliqua^b, Quercus coccifera^b, Rhamnus lycioides subsp. oleoides^b, Prasium majus^b, Rhamnus lycioides subsp. lycioides^a, Helianthemum spp^a, Juniperus phoenicea^c, Chamaerops humilis^c, Rhus tripartita^c, Lavandula multifida^c, Corydothymus capitatus^c, Erica multiflora^c, Calycotome spinosa^c, Fumana spp^c, Fagonia cretica, Slavia triloba, Asparagus acutifolius, Phlomis fruticosa, Satureja nervosa, Teucrium spp. Artemisia campestris.

2.2.2 In the arid bioclimatic zone (100 < R < 400 mm)

2.2.2.1 The Juniperus phoenicea series

This series occurs in the upper part of the arid zone with a mean rainfall usually above 250 mn; sometimes relicts of it may be found with an average rainfall as low as 160–200 mm in remote hills more or less isolated by a difficult accessibility and/or the lack of permanent water resources: The area occupied by this type of garrigue is of some 15 000 km², of which 5 500 are in Tunisia, 6 000 in Algeria, 2 000 in Morocco and 1500 in Libya.

The vegetation is of a sub-steppic type where forest relicts and/or remnants are more or less sparsely distributed low shrubs 0.5 m to 2.5 m high, often with a dominant under stratum of steppic species such as Stipa tenacissima, Lygeum spartum, Artemisia herba alba.

The main shrub species are:

Forest relicts or remnants: Juniperus phoenicea^c, Globularia alypum^b, Olea europea f. oleaster^b, Phagnalon saxatile^b, Helianthemum ellipticum^b, Withania frutescens^b, Onobrychis argentea^b, Fumana leavipes^a, Fumana ericoides^a, Fumana thymifolia^a, Helianthemum virgatum^a, Phagnalon rupestre^a, Phagnalon saxatile^a, Cistus libanotis^c, Genia microcephala^c, Rhus tripartitum^c, Launaea acanthoclada^c, Calycotome intermedia^c, Rosmarinus officinalis^c, Prasium majus^b, Rhamnus lycioides subsp. oleoides^b, Phagnalon rupestre^b, Periploca loevigata^b, Hippocrepis scabra^b, Coronilla minima^b, Teurcrium fruticans^a, Fumana loevipes^a, Pistacia atlantica^a, Helianthemum cinereum subsp. rubellum^a, Rosmarimus tournefortii^c, Thymus hirtus^c, Cistus salviifolius^c, Thymus spp., Rhamnus lycioides subsp. lycioides^c, Ziziphus lotus^c, Pituranthos scoparius^c, Steppic species: Artemisia herba alba^b, Argyrolobium uniflorum^b, Helianthemum lippii^b, Helianthemum hirtum^b, Artemisia campestris^c, Echiochilon fruticosum^b, Gymnocarpos decander^b, Helianthemum kahiricum^b, Atractylis serratuloides^a, Anabasis oropediorum^b.

2.2.2.2 The white sage series

This series comprises a complex group of plant communities usually dominated by Artemisia herba alba, which occurs on silty soils between the isohyets of 100 and 400 mm from SE Spain to Southern Russia throughout all Northern Africa and the Near East. The main shrub species are chamaephytes 0.10 to 0.50 m high, with densities of 1000 to 40 000 undershrubs per hectare and a ground cover from 5 to 40% depending on aridity and depletion status.

Artemisia herba alba^b, Anabasis oropediorum^b, Helianthemum kahiricum^b, Helianthemum virgatum^a, Salsola vermiculata var. villosa^a, Atractylis serratuloides^c, Atractylis phaeolepis^c, Ormenis africana^c, Astragalus armatus subsp. Trabacanthoides^c, Kikxia aegyptiaca^c, Anabasis aphylla, Helianthemum hirtum^b, Gymnocarpos decander^b, Argania sideroxylon (SW Morocco only)^b, Noaea mucronata^a, Acacia gummifera (SW Morocco only)^a, Hammada scoparia^c, Farsetia aegypticaca^c, Herniaria fontanesii^c, Asparagus albus, Atractylis humilis.

2.2.2.3 The field sage series

This series comprises various types of steppes developed on sandy soils throughout the same geographic zone as the white sage series.

The main shrubs are:

Artemisia campestris^c, Echiochilon fruticosum^b, Argyrolobium uniflorum^b, Helianthemum kahiricum^b, Helianthemum hirtum^b, Thymelaea microphylla^a, Rhantherium suaveolens^a, Salsola vermiculata subsp. villosa^a, Suaeda mollis^a, Artemisia monosperma^a, Ziziphus lotus^c, Nolletia chrysocomoidos^c, Thymelaea hirsuta, Ononix natrix subsp. falcata, Pergularia tomentosa, Helianthemum lippii^b, Polygonum equisetiforme^c, Salsola venniculata subsp. Brevifolia^b, Gymnocarpos decander^b, Argania sideroxylon (SW Morocco only)^b, Lygos raetam^a, Traganum nudatum^a, Acacia gummifera (SW Morocco only)^a, Hammada schmittiana^c, Calycotome intermedia^c, Atractylis serratuloides^c, Pitutanthos tortuosus^a, Pitutanthos rholfsianus^a, Nolletia chrysocomoides.

2.2.3 The desert bioclimatic zone

The desert (Saharan) bioclimatic zone is characterized by an average annual rainfall below 100 mm. However permanent pastures are only found between the isohyets of 50 and 100 m/m. Below the 50 mm isohyet, pasture production is extremely irregular and can be used only for a few weeks every so many years, consecutive to unusual rains or floodings.

They are located along the stream network and in depressions, and account for very little in the overall pasture production of the countries examined in this paper. The rangeland lying between the 50–100 mm isohyets, on the contrary, are used fairly regularly, at least during the springtime. Their acreage is very large (500 000 km²); they probably account for about 10% of the pasture production of the countries concerned in the present study.

2.2.3.1 Sandy steppes

The main shrub species are the following:

Acacia raddiana^b1, Helianthemum brachypodum^b, Traganum mudatum^b, Maerua crassifolia^b1, Moltkia ciliata^b, Genista saharae^a, Calligonum azel^a, Hammada schmittiana^c, Calligonum comosum^b, Cornulaca monacantha^b, Hedysarum argentatum (SW Morocco only)^b, Balanites aegyptiaca^b, Lygos raetam^a, Ephedra alata^a, Calligonum arich^b1, Euphorbia guyoniana.

¹Very rare.

2.2.3.2 Steppes on shallow and/or silty soils

Anthyllis sericea subsp. henoniana^b, Gymnocarpos decander^b, Traganum nudatum^b, Zilla macroptera^a, Salsola sieberi^a, Hammada scoparia^c, Randonia africana^c, Ziziphus lotus subsp. Saharae^c, Asteriscus graveolens^c, Astragalus pseudotrigonus^c, Anabasis articulata^c, Oudneya africana, Waronia saharae, Hyosciamus muticus, Pergularia tomentosa, Anabais aretioides, Helianthemum kahiricum^b, Nucularia perrini^b, Zilla spinosa^a, Artemisia judaica^a, Salsola tetragona^a, Crotalaria saharae^a, Antirrhinum ramosissimum^c, Anvillea radiata^a, Anabasis articulata^c, Alhagi maurorum^c, Salsola tetragona^a, Fagonia microphylla, Anabasis aretioides, Pulicaria crispa, Salsola baryosma, Solenostemma argel.

2.2.4 The halophytic steppes

The halophytic steppes cover huge areas in Algeria, Tunisia, Libya and Egypt. These pastures are mainly used by camels throughout the year but also by small stock during the dry season. The dominant species are mainly crassulescent chamaephytes of the Chenopodiaceae family.

Atriplex halimus^b, Atripley glauca^b, Atriplex malvana^a, Atriplex coriacea^a, Suaedea fruticosa^a, Suaedea brevifolia^a, Suaedea mollis^a, Salicornia arabica^c, Limoniastum monopetalum, Limoniastum guyonianum, Traganum nudatum^b, A triplex mollis^a, Atriplex portulaccoides^a, Salsola vermiculata var. villosa^a, Salsola sieberi^a, Salsola tetrandra^c, Arthrocnemum indicum^c, Salicornia fruticosa^c, Inula crythmoides^c, Halocnemum strobilaceum.

3. Introduced species

A good many browse species have been introduced from various parts of the world to northern Africa for the past 100 years and more. In many instances, however, these introductions were made for other purposes than browse; such is the case of the American vine stocks (Vitis rupestris, V. berlandieri, V. riparia and their hybrids), which turned out to be excellent browse and soil protecting species, although they have rarely been used to this aim. These were introduced some 120 years ago during the phylloxera crisis in Europe, and later for the expansion of vineyards over some 440 000 hectares. This is also the case of the mulberry introduced by the Arabs during the middle ages for silk production, as well as for its fruits. Morus alba is an excellent browse species as we shall see further; but it has hardly been used as such, at least for livestock feeding. The same thing applies to the olive, native to the country, but cultivated since some 2200 to 2500 years when the country was under Phoenician, then Carthaginian influence.

Cacti were introduced from Central America to Spain during Colombus's second expedition (1494–96) at the end of the 15th century. They were taken to North Africa by the Moors when these, a few years later, were finally expelled from Spain at the end of the 15th century (Monjauze et Le Houérou,1965). Cacti (mainly Opuntia,ficus-indica) have thus been in North Africa for nearly 400 years and have become naturalized and multiplied by birds, especially in cliffs and other inaccessible places in the semi-arid and subhumid bioclimatic zones. They were cultivated for their fruits and used as fences. Their deliberate use as fodder is fairly recent, dating probably from less than 80 years ago. Other browse species were successfully introduced at the beginning of the present century.

Gleiditschia triacanthos, the common honey locust was imported from the USA as a roadside ornamental tree, the pods of which are sometimes used as livestock feed. Honey locust is well adapted to the semi-arid and subhumid bioclimatic zone (400 < R < 800 mm) and can withstand very cold winters and relatively high elevations, probably up to 2000 m a.s.l.

Prosopis juliflora and P. chilensis were probably introduced from central and southern America between the first and second world wars and were well adapted to the arid zone below 600–800 m, i.e. wherever the mean minimum temperature of January is above +2°C.

Atriplex nummularia was introduced from Australia to Tunisia some 80 years ago as a browse species for arid and semi-arid zones. The first plantations established and used by farmers in Tunisia date back some 40 years. Its extension in the arid zone is much more recent and was not implemented before the late 1960's (Franclet and Le Houérou,1971). Other species of Atriplex were introduced during the 1960's: A. canescens, A. vesicaria, A. semibaccata etc. (see Franclet and Le Houérou, op. cit.). Another promising recent introduction (Le Houérou, 1965) is Chenopodium auricomum from central Australia.

Australian wattles (phyllodinous Acacias) were introduced especially in Morocco, Libya and Tunisia for the purpose of coastal sand dune fixation, in the 1920's and 1930's. This is especially the case of Acacia cyanophylla planted over many tens of thousands of hectares in these three countries. This species turned out to be an excellent browse. Other wattles were introduced more recently in the 1960s, especially A. salicina, A. ligulata and A. victoriae, which are very successful in the arid zone since they grow and produce in areas receiving as little as 150 mm of annual rainfall.

A very successful introduction in the semi-arid zone is Medicago arborea from the Greek islands. The same applies to the Russian olive, Eleagnus angustfolia, introduced from the Near East.

One should also report that some species that raised great hopes for the arid zone produced rather mediocre results; two outstanding examples are the Kachia or Maireana (as they should now be called), this is also the case of the mulga A. aneura, and of A triplex semibaccata.

Some native species have also been planted, although usually not for browsing purposes. Such as the olive, the carob, the ash (Fraxinus oxyphylla), the coronilla (Coronilla glauca, C. emeroides, C. valentina). The case of the olive deserves some comments as the cultivated olive occupies some 3 million hectares or 100 million trees in North Africa today. Leaves and twigs from pruning are fed to livestock while branches are transformed into charcoal. As the average annual production per tree is of the order of 10 kg of DM of leaves and twigs, the total production for the region is equivalent to the annual dietary needs of some 0.8 million sheep.

Other native browse species successfully developed in plantations in the arid zone are Calligonum comosum and Periploca laevigata.

The principal ecological requirements of cultivated browse species in Northern Africa are shown in Table 6.

Table 6. Ecological requirements of some artificially established browse species in North Africa.

From the above table it is obvious that with the available plant material, of some 35 species, one can meet most of the ecological conditions prevailing in northern Africa, except in the desert where rainfall does not reach the 100 m mark.

4. Forage value

The forage value of any consumed plants is the result of two main components:

a) its palatability and voluntary intake by livestock;
b) its nutritive value i.e. chemical composition and digestibility.

4.1 Palatability

4.1.1 General

Palatability is a very complex notion, very difficult to generalize as it is linked to many variables in time and space; some of these variables are linked to the plant, others to the animal while a third category depends on various environmental factors. For a given species palatability for a given type of animal varies with the phenological stage, the organ concerned and the season.

Moreover in almost any population, either natural or planted, of a given species of browse, there are various degrees of palatability from one plant individual to the net, ranging from highly palatable to poorly palatable. Palatability also depends on the relative abundance of the species under consideration on the rangeland; all other conditions being equal the palatability of a given taxon is inversely related to its abundance on the range, except for a few species which are specially relished in all circumstances ("ice-cream species" in American range management jargon).

Besides these "internal" plant factors there are also "internal" animal factors which are: species and race of livestock, age, feeding habits, physiological and health status, nutritional status (an undernourished and hungry animal is less selective than a well-fed beast). As a rule, the content of crude fibre in a forage plays an important role in its selection by livestock. Forages with a high fibre content are usually better accepted by cattle than by sheep and goats; but this, in turn, depends on the proportions of the various components of fibre: cellulose, hemicelluloses, acid detergent fibre (lignocellulose), neutral detergent fibre (cellwalls) etc. But there are many other criteria for selection such as organoleptic qualities of the forage; the latter have hardly been explored in research. Mineral content may also be an important factor either limiting or favourable (usually limiting in low rainfall areas and favourable in high rainfall areas, when silica-free minerals are concerned). Finally, the overall balance of the diet plays a major role in forage selection. It has been shown (Skouri, 1975) that the amount of highly fibrous material ingested depends to a large extent on the amount of protein in the overall diet. Besides these intrinsic or internal plants and animal factors there are also extrinsic or environmental criteria for selection.

Some of those have been referred to above, such as the relative abundance of a given taxon on the range and the botanical composition of the forage available. The palatability of a given taxon of browse for a given type of animal depends to a large extent on the plant community or association in which this taxon is being browsed, since most of them occur in many plant associations under various ecological and bioclimatic conditions. It is a matter of competitivity or rather of "demand and supply". This is why a given species may be differently rated in the various plant communities (see paragraphs 2.2.1 and 2.2.2 above).

This fact has been rightly stressed by Sarson and Salmon (1976). As a rule the palatability of a given taxon would increase with environmental aridity. There are, however, as usual, exceptions to this rule, perhaps due to unnoticed (by man) ecotypical or biochemical differences between various populations of a given taxon. An example of the first (normal) situation is Traganum nudatum more heavily browsed in the desert bioclimatic zone than in the arid zone. An example of the exception is Moricandia nitens (=M. suffraticosa) browsed in the arid zone of Libya, Morocco and Tunisia and ignored in the northern Sahara of Algeria. For these reasons an assessment of palatability of browse species is always to some degree subjective.

It is, however, possible to reach some degree of objectivity in the rating of palatability for a given consumption experiment with a given plant material and a given group of animals. This may be achieved by measuring the rate of use of the forage available. However the result of a given experiment can hardly be extrapolated and generalized. In Northern Africa Le Houérou (1962, 1965), Le Houérou and Ionesco (1971), Sarson and Salmon (1976) have proposed a scale of rating with 6 classes.

Class	Rating	Rate  ( =Biomass consumed )             Biomass Consumable
1. Highly palatable	HP	90–100%
2. Very palatable	VP	65–90
3. Palatable	P	45–65
4. Fairly palatable	FP	10–45
5. Occasionally palatable or   poorly palatable	PP	1–10
6 Not palatable		0–1

Loiseau and Sebillota (1972) used a 5 grade rating and Ziani. (1970) a 4 grade rating.

In the present study we shall consider four classes only:

b)  More or less palatable according to circumstances (classes 2-4).

c)  Poorly palatable (class 5)

d)  Not palatable (class 6)

There are over 600 ligneous species in North Africa. Some are rare or localised endemics which play practically no role in animal nutrition at the regional scale. The following list of some 300 species is restricted to rather common species or to uncommon species which are of high value and would deserve attention as to their agronomic potentials. The bulk of browse, however, is made out of a limited number of species (less than 100), whereas other species, of great potential interest, play a minor role because of their relative rarity.

4.1.2 Rating of palatability of native ligneous species.

Table 7

Captions to the table

Column 2: Bioclimatic zones

D = Desert (mean rainfall < 100 mm)
A = Arid (100 < R < 400)
AI = Lower arid (100 < R < 300)
AS = Upper arid (300 < R < 400)
SA = Semi-arid (400 < R < 600)
SH = Subhumid (600 < R < 800)
H = Humid (800 < R < 1200)
PH = Perhumid (1200 < R)

Column 3

F = Leaves
R = Twigs
Fl = Flowers
F = Fruits

Column 4

HP = Highly palatable
P = Palatable
PP = Poorly palatable or occasionally palatable
NP = Non palatable
To = Toxic

Column 5

Bov = usually browsed by bovines, cattle
Ov = usually browsed by ovines, sheep
Cap = usually browsed by caprines, goats
Ch = usually browsed by camelines
                              (dromedaries)

(Bov) occasionally browsed by bovines
(Ov) occasionally browsed by ovines
(Cap) occasionally browsed by caprines
(Ch) occasionally browsed by camelines

1. Cistus species: C. albidus, C. crispus, C. incanus, C. ladaniferus, C. laurifolius, C. libanotis, C. monspeliensis, C. parvifolius, C. polymorphus, C. salicifolius, C. salvifolius, C. sericeus, C. varius, C. villosus.
   
2. Euphorbia, ligneous species: E. balsamifera (Maroc), E. baumierana (Maroc), E. bivonae, E. characias, E. dendroides, E. echinus (Maroc), E. guyoniana, E. regisjubae (Maroc), E. resinifera (Maroc).
   
3. Common Genista species: G. cacanthoclada (Libye), G. aspalathoides, G. caballeroi (Maroc), G. cephalantha, G. cinerea, G. demnatensis (Maroc), G. erioclada, G. ferox, G. florida, G. microcephala, G. myriantha (Maroc), G. numidica, G. quadriflora, G. retramoides, G. saharae, G. spartioides, G. tricuspidata, G. ulicina, G. umbellata, G. vespres.

4.2 Voluntary intake

There are, unfortunately, very few experimental data on this subject in the region. El Hamrouni and Sarson (1976) working with sheep on a purely browse diet in the maquis of northern Tunisia found that intake of browse is closely related to the liveweight of the animals. The relation found was:

           Y = 37.85 X +43.2

where Y = grammes of DM ingested daily

X = liveweight of animal in kg.

This corresponds to a daily consumption of 3.8 kg of DM per 100 kg liveweight. Working with goats, Sarson in central Tunisia and Novikoff in Southern Tunisia, found a daily consumption of 57 and 60 g of DM per kg of liveweight respectively. Cattle are not supposed to ingest daily more than 2.9 kg of DM per 100 kg of liveweight; in fact this classical figure seems too low, both in northern Africa and in the tropics where consumptions of 3.0 to 3.5 kg/100 kg liveweight have been reported.

In summary, cattle, sheep and goats can consume up to 3, 3.8 and 6.0% of their body weight in dry matter daily. Camels can ingest daily 2.5 to 3.0% of their body weight in dry matter (a figure I have computed from various sources quoted by Ortiz and Mukassa, 1979). However, like goats and unlike sheep and cattle they can thrive on a permanent basis on a pure browse diet.

Sarson and Salmon (1978) have clearly and satisfactorily explained why cattle and sheep cannot meet their nutritional needs on pure browse diet whereas goats do. This is shown in the following table where the needs and possibilities of ingestions per 100 kg of liveweight are given for cattle, sheep and goats.

Maintenance needs/100 kg liveweight	Cattle	Sheep	Goats
A FU	1.9.	1.33	1.13
B Quantity DM ingestible (kg) per day	2.9	3.8	6.0
Minimum ratio FU/kg DM to meet maintenance needs	0.65	0.35	0.19

As nutritional value of browse is of the order of 0.25–0.40 FU per kg of DM, according to Sarson and Salmon, browse alone cannot ensure the maintenance requirements of cattle (0.65 FU/kg DM); browse can ensure maintenance of sheep (0.35 FU/kg DM) but does not allow production; with goats maintenance and production may be provided on a pure browse diet (0.19 FU/kg DM).

This explains why only goats, camels and a few wild herbivores can survive on depleted rangelands, as often occur in arid zones where browse often constitutes the only feed available for long periods of time. This is also why goats and camels are much less affected by catastrophic droughts than sheep and cattle, as experienced in the Sahel in 1969-73 and in East Africa in 1971-74.

4.3 Chemical composition and nutritive value

Table 8 (1-32) shows the data from 323 analyses of some 110 browse species belonging to 78 genera and 30 families. Eighty are native species while 30 are exotic but extensively used in the region. Basic chemical analyses are:

b)  Ashes or total minerals obtained by calcination in furnace at 600°C.

c)  Silica: part of ashes non soluble in HC1.

d)  Crude protein: N Kjeldahl x 6.25.

e)  Crude fibre residue from reaction by concentrated triacid solution under heat.

f)  Crude fat: Ether extract.

g)  Macro-elements: P, Ca, K, Na measured either by complexometry or spectrophotometry.

h)  Nitrogen free extract: difference 100– (CP + CF + Ash + FAT)

Nutritive value has been estimated from these data by computation using the so called "Dutch tables" (Dijkstra, 1957) for green legumes. One table gives the net energy value knowing the contents of the forage in ashes and crude fibre; another table provides the figure for digestible protein knowing the crude protein. The tables established for legumes show a content in net energy about 10% lower than those established for grasses for a given combination of CF and ash. Digestible protein amounts to approximately: 0.93 CP –3.52 (Demarquilly et Weiss, 1970). All data are expressed in percentage of dry matter.

I have not used the evaluation of nutritive value made by the various authors of the data utilized, except those of Piccioni, which are separated and given for the sake of reference and comparison. This is because there are very large differences of interpretation between the various authors due to their use of various ways of estimating forage value. These differences may vary from 50% less to 200% more, or even greater than that, for a given combination of raw data in chemical analyses. We shall come back to this point further on. In this situation the comparison of such data is meaningless.

In order to render the data compatible and comparable I have only used the raw data from chemical analyses, i.e. DM, CP, CF, ash, ENA, FAT, provided by the authors of the various publications consulted. I have then computed the forage value from Dijkstra tables for green legumes. These tables concern forages having 18-42% CF and 5-25% ash, and 8-25% ash, and 8-25% CP. Extrapolation of the tables were made for the cases not fitting these frames, which amount to a relatively limited number.

Source	Ash	CP	CF	Fat	Dig. Coef. OM	FU Kg DM	MJ kg DM
Wilson	29.5	20.6	–	–	68.8	0.40	2.75
El Hamrouni and Sarson	30.4	22.0	10.0	3.3	52.8	0.28	1.57
Le Houérou, Extrap. of Dijkstra's tables	26.1	19.3	13.3	3.1	–	0.72	5.00

This method, I am fully aware, is subject to criticism; it does, however, have the great advantage of making the evaluations of fodder value comparable for all analyses examined in this paper, which would not be the case otherwise, as mentioned before. In other words the evaluations made in the present paper are not correct and accurate in absolute value terms (they cannot be so, anyway as long as there are not detailed and numerous digestibility experiments) but they are comparable. Had I only reported the various authors' evaluations, the data would have been neither correct nor comparable.

Some authors have, probably very rightly, expressed serious doubts about the validity of using the "Dutch tables" (established from classical temperate climate fodders) for Mediterranen or tropical fodders, and especially browse. The facts are that the too few digestibility experiments available in the literature (in vivo and in vitro) show great variability in the digestibility of browse (30 to 72% of the organic matter).

Geri and Sottini (1970) working on browse species of the Mediterranean maquis in Sardinia found a regression fitting the following equation: DC = 57.49 – 0.232 CF– 0.725 Fat where DC = in vitro digestibility coefficient of the organic matter CF = crude fibre; Fat = Crude fat. This predictive equation results in Apparent Digestibility Coefficients of Organic Matter usually below 50% (with an average 25% CF and 4.5% fat).

However several authors have found Apparent Digestibility Coefficients that are sometimes much higher than 50% for browse, using either in vivo or in vitro method. For instance: Rose-Innes (1976): 54–70% (O.M.), Ghadaki Van Soest et al (1974): 52-59% (O.M.), Wilson (1977): 29-69% (O.M.), Leight, Wilson et al (1978): 29-62% (D.M.).

E1 Hamrouni and Sarson have calculated the nutritive value of some 110 species of browse in Tunisia using Geri and Sottini's regression for the determination of the digestibility coefficients and Breirem's equation for the evaluation of Net Energy:

NE =236 D OM –1.2 (OM-D OM)
                     1650

where NE = Net Energy in Feed units per kg of DM
          OM = Organic Matter
          DOM = Digestible Organic Matter (OM x AD
                       COM)
         ADCOM = Apparent Digestibility Coefficient of
                          Organic Matter

The values found by El Hamrouni and Sarson seem very low as compared to more classical figures. If, for instance we compare the Net Energy in A triplex nummularia leaves, using three different methods of evaluation, we obtain the following figures:

E1 Hamrouni and Sarson find an average 2.1 MJ (0.30 FU) per kg of DM over 110 species of browse in Tunisia; whereas using Dijkstra tables, we find 4.8 MJ (0.70 FU) over 110 species.

The use of the Dutch tables and their extrapolation result in Net Energy figures that seem much too high, whereas E1 Hamrouni and Sarson's figures would seem too low (although the data they reach for classical forages such as alfalfa or fescue are generally in agreement with the literature).

It may well be that E1 Hamrouni and Sarson's figures are the closer ones to the real values. I feel, however, that we need further evidence before this method of estimation of the feeding value of browse is generalized. It is even doubtful whether a generalized and valid predictive model could be found for all browse, given the large differences in chemical composition and digestibility between families and species.

Table 8 31 shows the average figures for the 323 analyses of the 110 species studied on North Africa; and for the sake of comparison, the average figures obtained by El Hamrouni and Sarson in Tunisia from 120 analyses of 112 species (48 natives + 64 exotics), the average figures for West tropical Africa are also shown (540 analyses over 105 species). The raw data of chemical analyses are very similar except for crude fibre which is markedly lower in West African browse, probably due to the fact that the data include a relatively important proportion of pods of legumes having a low fibre content.

The interpretations of the nutritive value differ very strongly due to the different methods of evaluation in the three cases. West African browse analyses were interpreted using the tables of Dijkstra for grasses. Whereas North African browse analyses were assessed by the use of Dijkstra's tables for legumes, in which the amount of energy for any given combination of ash and crude fibre is about 6% lower than for grasses. Therefore the average of 6.0 MJ per kg of dry matter in West African browse would have been about 5.7 MJ had one used the legumes table. This figure comes then closer to the one found for North African browse (4.9 MJ). Now if we interpret E1 Hamrouni and Sarson's raw data through the same table we find the average Net Energy value of 5.4 MJ/kg DM, which is halfway between the two other figures. If we interpret E1 Hamrouni and Sarson's data using the "Grasses table" of Dijkstra one finds a Net Energy value of 5,96 MJ/kg MD, i.e. 2.8 times higher than the one arrived at by these authors with the method described above. In other words the interpretation of chemical analyses for the evaluation of feed value of browse remains a field of high uncertainty and controversy.

The issues raised here seem of paramount importance; intake, digestibility and nutritive value of browse should be given a high priority in animal nutrition research in northern Africa as well as in the tropics, if one is to make a realistic and reliable assessment of the value of browse in the feeding of ruminants. The issue is one of development policy: should one develop or not the use of browse, through browse plantations for instance?

There is a dearth of data on trace elements and carotene content of North African browse. Aguer (1973) finds an average of 36 mg (17-74) of β carotene per kg of dry matter in the leaves of Atiiplex halimus and A. nummularica in summer in Tunisia; he suggests that the content in carotene may be one of the criteria for browsing preference between species and strains of Atriplex offered in cafeteria trials.

4.4 Apparent digestibility coefficient of organic matter in some North African browse species (average values)

Species	ADCOM	Method	Reference
Acacia aneura	45.6	in vivo	Woodman, 1942
Acacia aneura	44.0	In vitro	Wilson, 1974
Acacia pendula	45.0	in vitro	Wilson, 1977
Anthyllis vulneraria	57.3	in vivo	Piccioni, 1965
Arbustus unedo	75.0	in vivo	Schmidt-Burr, 1972
Arbutus unedo	55.7	in vitro	Geri-Sottini, 1970
Artemisia herba alba	69.1	in vivo	Ziani, 1966=
Artemisia herba alba	54.2	in vitro	Ghadaki, Van Soest and al, 1974
A triplex canescens	57.8	in vitro	Ghadaki; Van Soest and al, 1974
A triplex canescens	54.0	in vitro	Ghadaki, Van Soest and al, 1974
A triplex glaucca	67.9	in vivo	Schmidt-Burr, 1970
A triplex nummularia	68.8	in vitro	Wilson, 1977
A triplex nummularia	46.0	in vivo	Ben Ameur-Blomeyer, 1974
A triplex semibaccata	59.8	in vivo	Corriols, 1965
Atriplex vesicaria	52.9	in vitro	Wilson 1977
Ceratonia siliqua (Fr)	89.6	in vivo	Woodman, 1942
Erica arborea	40.1	in vitro	Geri-Sottini, 1970
Fraxinu oxyphylla	65.3	in vitro	Geri-Sottini, 1970
Myrtus communis	55.6	in vitro	Geri-Sottini, 1970
Noaea mucronata	42.3	in vitro	Ghadaki, Van Soest, 1974
Olea europaea oleaster	44.0	in vivo	Ben Ameur-Blomeyer, 1974
Opuntia ficus-indica	60.3	in vivo	Theriez, 1966
Phillyrea angustifolia	51.0	in vivo	Schmidt-Burr, 1972
Pistacica lentiscus	36.8	in vitro	Geri-Sottini, 1970
Pistacia lentiscus	38.8	in vivo	Derkaoui, 1977
Prosopts juliflora (Fr)	76.0	in vivo	Frays, 1925
Quercus ilex	40.9	in vitro	Geri-Sottini, 1970
Quercus pubescens	45.7	in vitro	Geri-Sottini, 1970
Spartium junceum	62.2	in vitro	Geri-Sottini, 1970

5. Productivity

5.1 General remarks

Browse productivity is not very well known for a number of reasons, one being the difficulty of accurately measuring consumable primary production. Secondary production, or in other words the primary production actually consumed as measured in terms of animal intake, is no easier to estimate, as browse plants are always mixed to a greater or lesser degree with other kinds of feed: grasses, legumes, forbs, etc. Nevertheless, measuring productivity in artificial plantations presents no problem, owing to their homogenous nature in terms of age, phenology, size, spacing, soil, etc.

On the other hand, the various productivity values are spread out over a wide range, according to ecological conditions and, above all, mean annual rainfall and soil quality. In North Africa, natural browse ecosystems are limited to uncultivated areas, or in other words land which is too arid or too poor to be cultivated. There is thus a certain homogeneity in the soil conditions in so far as these are always poor for one reason or another, such as too much or too little water, shallowness, toxicity, etc. For this reason we shall examine productivity over broad ecological zones basically determined by average rainfall. Nonetheless, in any given range of rainfall productivity also varies according to the botanical composition and vegetation structure of the browse community. Botanical composition and vegetation structure largely depend, in their turn, on management practices and in particular on the degree of degradation and the resilience of the plant communities resulting from these practices. In any given ecological zone productivity often varies by a factor of 1 to 5, and sometimes more, in accordance with these factors.

Moreover, Liacos and Moulopoulos (1967) and Papanastasis and Liacos (1980) have demonstrated that great differences in palatability and productivity which occur between biotypes within a given population of Quercus coccifera in Greece. Elsewhere, these two aspects have not been much explored, although qualitative observation of both native and artificial plantations demonstrates clearly that palatability and productivity range widely within practically any population of any browse species. For all these reasons, our evaluation of browse productivity can only be approximative.

5.2 Production of browse in the semi-arid and humid bioclimatic zones

As mentioned above (Section 2.2), rangeland pasture in the semi-arid and humid zones (P>400mm) is made up almost entirely of forests or degraded forests (matorral), and specially of the latter. Production figures vary from under 200 to over 1500 kg of consumable feedstuffs per ha and per year, according to vegetation types and rangeland condition. (Le Houérou, 1971 and 1975; Sarson and Salmon, 1977). These figures are comparable to those found for similar types of vegetation on the northern shores of the Mediterranean (le Houérou, 1971, 1980).

Some types of vegetation are virtually unproductive, even in areas of heavy rainfall and on good quality soils. Such is the case for dense populations of Cistus spp. and, in particular, C. monspeliensis or Pteridium aquillnum (it should be mentioned that these types of vegetation are the result of uncontrolled fires and overstocking). (Le Houérou, 1973, 1980).

Other types of vegetation provide relatively high production around 1000-1500 of DM a year of edible (or 550,000 to 775,000 kcal of net energy, i.e. 2300 to 3240 MJ). Such is the case for matorrals dominated by Quercus ile, Q. pyrenaica, Q. suber, Q. faginea, Olea europea, Phillyrea angustifolia, etc.

Some species are pollarded, and the branches and twigs given as feed to livestock in the farms and villages. This practice is earned out with the ash, Fraxinus oxyphylla, the elm, Ulmus capestris, the oaks, Q. pyrenaica, Q. suber, Q. faginea, the arbutus, Arbutus unedo, the olive 0. eruopa sativa, and the carob, Cerationia siliqua.

Average productivity varies according to annual rainfall (Le Houérou and Hoste, 1977); production of natural browse on rangeland alone is about 1 to 1.5 kg of DM actually consumed per ha and per mm of rainfall, taking into account that only 50% of consumable biomass (leaves, twigs, flowers and fruit) is in fact consumed owing to problems of access: high branches out of reach of the animals and the inner production of bushes with an introverted habit. Thus, in the semi-arid zone, production is 500–600 kg of DM consumed, in the subhumid zone 600–800, in the humid zone 800–1200 kg, and in hyperhumid areas 1220–2000 kg.

Artificial browse plantations have a much higher production as they are planted on better soils, and because they are better managed and maintained. The carob produces from 40 to 80 kg of fruit per tree per year. An adult plantation of 600 trees ha thus produces 6000 kg of pods, or 8.8x10 kcal/ha/ year in terms of net energy, enough to satisfy. The energy needs of 20 adult sheep for a year.

The production of the false acacia (Gleditschia triacanthos) is comparable to that of the carob, except that the tree starts producing earlier (5–10 years old instead of 10–15). Spineless cacti (especially Opuntia ficus indica var. inermis), are planted mainly in the arid and semi-arid zones. In the semi-arid zone, the production of well-managed adult plantations varies from 50 to 200 t of green matter per ha per year, with an average of 80 t, that is 7000 kg of DM/ha/year (Monjauze and Le Houérou,1965). As there are more than 200,000 ha of cacti in the semi-arid and subhumid zones, the total production is this in the region of 1.4 million t of DM.

Pruning olive trees produces about 10 kg of DM of leaves and twigs per year, per tree; as there are about 100 million cultivated olive trees in North Africa, this by-product represents the yearly energy needs of some 0.8 to 1 million sheep, or 3.3% of the forage fodder available for the North African herd.

Plantations of tree Lucerne (Medicago arborea) can produce around 3500 kg of DM/ha year¹, in the areas of annual rainfall of 350–450 mm, or 4 million kcal of net energy and 350 kg of DP per ha per year (E1 Hamrouni and Sarson, 1974).

¹These figures concern only the productivity of browse plants; the additional production of the grass stratum is roughly 50% of that of bushes.

Plantations of tree coronilla (C. glauca and C. valentina) produce nearly as much fodder of the same quality; as free Lucerne,.i.e. with a feed value nearly as rich as common lucerne.

Plantations of salt bushes, especially Atriplex nummularia and A. halimus, produce nearly the same quantity of fodder as tree Lucerne, and of a quality nearly as rich and digestible (Le Houérrou, 1969; Franclet and Le Houérou 1971; Aguer, 1973). In other words, the average bush, whether it be tree Lucerne coronilla, or Atriplex, produces on average 1 to 3 kg of edible DM a year in the semi-arid bioclimatic zone (350-600 MM) according to the density of the plantation (1000 to 3500 bushes/ha) and the management methods used. The production figures quoted above were obtained by harvesting, carried out simultaneously by men and sheep. These plantations of tree Lucerne coronilla and saltbush, which were 35–40 years old in 1980, are still being used at Ksar Tyr-Montarnaud, in Tunisia.

5.3 In the arid zone

In the arid zone the production of chamaephyte shrubs, such as Artemisia herba alba, Rhanterium suaveolens, Gymnocarpos decander, Helianthemum lippii, Helianthemum kahiricum, Echiochilon fruticosum, Salsola vermiculata and Anabasis orpediorum can represent the greater part of fodder production, i.e. 60 to 80%, or 50–90%, expressed in kg of DM/ha/year, according to Le Houérou et al (1974) and Floret and Pontanier (1978). In Algeria, Le Houérou et al found an average figure of 74% in the Hodna area. For southern Tunisia, the figure is 69% for a period of 7 years, on different types of pastures.

This proportion varies according to seasonal rainfall distribution, the nature of the soil and the topography, the condition of the rangeland and its degree of degradation. In Algeria, for example, Le Houérou et al found that shrubs made up 80% of production on pastures on which grazing had been prohibited for 5 years, consisting of Artemisia herba alba and Neaea mucronata, whilst on pastures with similar plant communities subject to continuous grazing the proportion fell to 64%. Grazing experiments in the same region have shown that the proportion of shrubs in fodder production was inversely proportional to the stocking rate:
64% for 1 sheep/ha
74% for 1 sheep/2ha
76% for 1 sheep/4ha
78% for 1 sheep/6ha
80% 1 sheep/pasture banned for grazing.

The total production of the steppes of North Africa varies from 9.5 to 5 kg of DM/ha/year for each mm of rainfall. The average is 2 kg of edible DM/ha for each mm of rainfall (Le Houérou, 1975; Le Houérou and Hoste, 1977).

Floret and Pontanier (1978) found a figure of 4.6 kg of DM/ha/year/mm on pasture in good condition consisting of Rhanterium suaveolens and Stipa lagascae near Gabes: this was based on observations made over a period of 7 years. In another relatively degraded site, more typical of the current state of grazing land they found 2 kg of DM/ha/year/mm during the same period and in the same area. On a poorer pasture made up of Anarrhinum brevifolium and Zygophyllum album growing on a gypsum layer in the same region, they found 1.18 kg of DM/ha/year/mm over the same period. These figures coincide perfectly with the predictions of Le Houérou and Hoste. Given an average shrub production of 70% in the overall production of the arid zones rangelands, the average quantity produced would be 1.4 kg of DM/year/mm. The total shrub production of the arid zone rangelands in North Africa would therefore be:
(3510⁶ha) (1.4 kg 250 mm) =12.25 10⁶ tonnes of DM/year.
Assuming utilization of 50%, this represents the fodder needs of 12 million adult sheep.

The desert pastures cover some 52 million ha; their bushes produce:
1.4 kg 75 mm 52 10⁶ ha = 5.4 10⁶ tonnes DM;
That is, the equivalent of the fodder needs of a further 5.4 million adult sheep.

The bushes of the arid and desert zones rangelands are thus capable of nourishing 17 million adult sheep, or more than half (53%) the ovine population of the region.

6. Management

Nearly all the North African rangeland belongs to the community or the state, and grazing rights are adjudicated according to the traditional legal system. Hardly any rangeland is individually owned, fenced in for the use of the owner alone. In such a situation it is in the interest of each user to take maximum advantage of the common resources, without the relationship being reciprocal. This system which effectively amounts to looting can only lead to exhaustion of the rangelands and a decline in productivity.

Moreover, since cultivated land is, on the contrary, worked individually, either according to traditional rights of tenure or as private property, and since population pressure is increasing very fast, all the cultivable rangeland either has been or is in the process of being subject to clearance and cultivation. Rangeland thus tends to be limited to land with soil which is either too stony or too shallow, arid, or salty to be cultivated.

Nevertheless, over the last 50 years many experiments and a limited number of large-scale development plans have demonstrated that most of the rangeland could be restored and its production doubled, tripled or even quadrupled under good management conditions. Good management means:

b)  practising rotational grazing in order to allow recovery of palatable species;

c)  using various improvements such as water conservation, erosion control, the setting up of watering points for livestock, and reseeding of pastures;

d)  building up fodder reserves, such as plantations of browse fodder.

These techniques are all well known and applied in other parts of the world, and it has been demonstrated that they are both applicable and economically viable in North Africa as well. The establishment and management of browse plantations was the subject of a separate study made by the author of this report, so the subject will not be treated here.

Caption to table 8

The symbols used in the tables below read as follows:
Column 2
FJ = Young leaves                            Fv = green leaves
R = twigs                                           FS = dry leaves
Fr = Fruits                                         FrV = green (immature) fruits
FrS = dry fruits                                  Ml = meal
G =Seeds                                         GM = mature seeds,
GS = dry seeds
F = leaves                                       Column 3
FM = mature leaves                         Alg. = Algeria
FJ = young twigs, sprouts                Lib. = Libya
FrJ = young fruits                              Eg. = Egypt
Fl = flowers                                        Mar. = Morocco
FJ = immature seeds                        Tun. = Tunisia
Am = Kernel                                      N.A. = North Africa

Column 4

month – year: 06 70 = June 1970
>54: no date but before 1954 where the data were published
SS = dry seasons
SSF = cool dry season (Nov. –Feb. in Sahel)
SSC = hot dry season (March–June, in the Sahel)
SH =rainy season

Column 5

Dry matter (DM) in % of fresh matter
Column  6  Crude protein (CP) in % DM
Column  7  Crude fibre (CF) in % DM
Column  8  Crude fat in % DM
Column  9  Nitrogen free extract (NFE) in % of dry matter
Column  10  Digestible protein (DP) in % of dry matter
Column  11  Net energy (NE) in MJ per kg of DM
Column 12 Nutritional ratio DP (g/kg DM)
                                            FU (kg DM)
Column 13  Ash (total minerals) in % DM
Column 14  Silica in % DM
Column 15–19 Macro-elements in % DM
Column 20  Reference (see no in bibliography)

7. Conclusions

It can be said that browse plants form the basis of animal feeds in North Africa. They constitute 60 to 70% of rangeland production, which in its turn, represents some 60% of animal feed resources. It can thus be estimated that browse plants represent around 40% of the total animal feeds available in the region.

The different browse ecosystems of the region have been examined and briefly described as far as their structure, botany, dynamics and productivity are concerned. The annual productivity of browse plants is about 1.5 kg of DM/ha/mm of annual rainfall, of which roughly 50% is actually consumed. About 35 types of browse plants, most of them exotic, have been tried out and used in the region with favourable results, which proves that it is possible to establish man-made browse reserves under nearly all ecologic conditions with the exception of pure desert. These man-made reserves represent some 0.5 million ha in the region.

The palatability, chemical composition and nutritive value of some 300 and 110 species respectively have been studied. Most browse production comes from a limited number of species (100). Chemical composition and nutritive value show that browse plants are rich in protein and mineral salts; they thus provide relatively good fodder in terms of net energy value. Browse intake depends on the species and breed of animal; sheep are able to take in 30–40% more DM per unit of weight than cattle, while goats can ingest twice as much; camels although able to feed solely on browse plants, take in less DM per unit of weight than cattle.

Unfortunately, owing to inadequate or insufficiently enforced legislation, browse grazing land is gradually diminishing not only in area owing to clearance for cultivation, but also owing to over utilization in productivity.

From the research point of view it is difficult to evaluate the nutritive value of browse accurately, because of the scarcity of data on digestibility intake and energy value. The percentage of carotene and trace elements in browse is an important factor in animal nutrition of the region, but there is scarcely any documentation on the subject. There are also very few trustworthy data on browse productivity.

The principal obstacles to an improvement in the situation lie in the fact that there is no adequate legislation covering the utilization of the rangeland of the region, of a kind which would ensure long-term sustained productivity. As long as rangeland resources are used communally while herds and flocks are private property, it is in the interest of each user to extract the maximum advantage from communal resources, without worrying about the long-term effects. Another bottleneck is the lack of local, experienced technicians to apply development plans, as well as inadequate administrative structures, which should consist of people competent enough to deal with the difficult and complicated problems involved in the management and development of rangelands.

References

Aguer D., (193). 'Compte-rendu d'un essai de pâturage d'Atriplex par des béliers â Bou-Rebia en Août-Septembre 1972'. Mimeo. Tunis. Institut national de la recherche agronomique de Tunisie.

Ben Ameur, M., Blomeyer, A. (1974) 'Composition chimique et valeur alimentaire des fourrages grossiers'. INRAT. Tunis.

Cordier G., (1047). 'De la composition de quelques produits fourragers tunisien et de leur valeur pour l'alimentation du mouton'. Annales Service Botanique et Agronomique de Tunisie, 20, 25–108.

Corriols, F., (1965). 'Essai d'adaptation de plantes fourragéres en Tunisie. Annales de llnstitut National de la Recherche Agronomique de Tunisie, 38, Tunis.

Curasson, M.G., (1958). 'Arbres arbustes, buissons et fourrages spontanes divers en région tropicale et subtropicale'. Rev. Elev. Méd. Vét. Pays Trop., VI, 2, 37–56; 91–112.

Derkaoui, M., (1977). 'Comportements alimentaires comparés des ovins et des caprins sur un parcours forestier de montagne (vallée de l'Azzem)'. Mem. 3e cycle. Rabat, Institut Agronomique et Vétérinaire Hassan II.

Delhaye, R., Le Houérou H.N., Sarson M., (1974). 'Amélioration des pâturages et de 1'élevage dans la région du Hodna'. AGS/DD/ALG/66/509 Rapp. Techn. no. 2. Rome, FAO.

El Aouini, M., Sarson M., (1976). 'Production primaire et valeur pastorale de certains types de maquis des Mogods non calcaires. Note de Rech. no. 12. Tunis. Institut National de Recherches Forestiésres de Tunisie.

El Hamrouni, A., Sarson M., (1974). 'Valeur alimentaire de certaines plantes spontanées ou introduites en Tunisie'. Mimeo. Note Rech. no. 2 Inst. Nat. Rech. Forest. Tunis.

El Hamrouni, A., Sarson M., (1076). 'Résultats d'un essai de charge sur une parcelle de Médicago arborea'. Note de Rech. no. 14, Inst. Nat. Rech. Forest. Tunis.

Floret, Ch., Pontanier R., (1978). 'Relations climat-sol-vegétation dans quelques formation végétales spontanées du Sud Tunisien'. Mimeo, CNRS/CEPE Montpellier, Paris, ORSTOM.

Foury, A., (1954). 'Plantes fourragères pour zones salées. Conférence Européenne des Herbages. Paris, Juin 1954 et La Terre Marocaine, 301, 447–454.

Franclet, A., Le Houérou H.N. (1971). 'Les Atriplex en Tunisie et en Afrique du Nord'. FO:SF/Tun 11, Rapp. Tech. no. 7, Rome FAO.

Fraps, G.S., (1925). 'Digestion experiments with cat byproducts and other feeds'. Texas Agric. Exp. Stn. Bull. p. 315.

Gaston, A., (1977). 'Survey for the development of the central wadi zone and gulf of Sirte. Grazing project, 3'. Agrostological report.Mimeo, Paris, IEMVT/GEFLI.

Geri, G., Sottini E., (1970). 'Composizione et digeribilità in vitro dei germogli a delle foglie di alcune piante della macchia inediterranea'. Firenze. Alimentazione Animale, 14.

Ghadaki, MD., et al (1974). 'Chemical composition and in vitro digestibility of some range forage species of Iran. 12th International Grassland Congress.

Le Houérou, H.N., (1962). 'Les pâturages naturels de la Tunisie aride et désertique'. Tunis. Institut des Sciences Economiques Appliquées.

Le Houérou, H.N., (1965). 'Improvements of natural pastures and fodder resources. Report to the Government of Libya'. EPTA Rep. no. 1979. AGP. Rome, FAO.

Le Houérou, H.N., (1073). 'Fire and vegetation in the Mediterranean basin. Proceedings Annual Tall Timbers Fire Ecology Conference, Tallahassee, Fa.

Le Houérou, H.N., (1974). 'Principles methods and techniques for range management and fodder production', AGP, Rome, FAO.

Le Houérou, H.N., (1975). 'The natural pastures of North Africa; types, production, productivity and development. Proceedings of Intern. Symposium on range inventory and mapping in tropical Africa, Bamako. Addis Ababa, International Livestock Centre for Africa.

Le Houérou, H.N., (1980a). 'The arid rangelands in Africa North of the Sahara'. In: Kaul and Thalen, eds.: Rangelands in arid zones: The Hague, Jung. In press.

Le Houérou, H.N., (1980b). `Impact of man and his animals on Mediterranean vegetation. In: Di Castri and Goodall eds. Mediterranean-type shrublands, Amsterdam, Elsevier.

Le Houérou, H.N., Haywood M., Claudin J., (1975). 'Etude phytoécologique du Hodna'. AGS: DP/ALG/66/509. Rome, FAO.

Le Houérou, H.N., Hoste C., (1977). 'Rangeland production and annual rainfall relations in the Mediterranean basin and in the African Sahelosudanian zone'. J. of Range Mgnt, 30, 3, 181-189.

Le Houérou, H.N., Ionesco T., (1973). 'Appétabilité des espéces végétales de la Tunisie steppique'. AG-TUN 71/ 525, Rome FAO.

Loiseau, P., Sebillotte M., (1972). 'Etude cartographique des pâturages du Maroc Oriental'. MARA, ERES, SCETCOP. Rabat, Paris.

Maignan, F., Salmon P., Sarson M., (1977). 'Approche methodologique pour 1'étude de la production des parcours forestiers'. Mimeo, FO: MOR 73/0 16, Rabat-Rome, FAO.

Monjauze, A., Le Houérou, H.N. (1965). 'Le role des Opuntia duns 1'économie agricole Nord Africaine'. Bull. Ec. Nat. Agric. Tunis, 8-9, 85-164 Tunis.

Nègre, R., (1974). 'Les pâturages de la région de Syrte (Libye): Projet de régénération'. Feddes Repertorium, Bd. 85, Heft 3, f, 185-243.

Ortiz, A. and Mukassa-Mugerwa, E. (1979). 'The camel: review of some aspects of the physiology productivity and diseases of the dromedary based on a literature review'. Addis Ababa, International Livestock Centre for Africa.

Papanastasis, V.P., Liacos, L.G. (1980). 'Productivity and management of Kermes oak brushlands for goats in Greece'. Addis Ababa, Proceedings Browse Symposium, International Livestock Centre for Africa.

Pellier, M., (1967). 'Note sur la composition chimique et la valeur alimentaire de quelques produits fourragers'. Rapp. Fonds Exper. et Rech. Agric., Execice 1956-57. Mimeo. Min. Agric. Tunis.

Piccioni, M., (1965). Dictionnaire des aliments pour les animaux. Bologna, Edagricola.

Rodin, L.E., et al (1970). Etudes géobotaniques des pâturages du secteur Ouest du département du Medea (Algerie). Leningrad Naouka.

Sarson, M., Salmon P., (1076). 'Appétabilité de certaines espéces vivaces spontanées ou introduites au Maroc et en Afrique du Nord'. Mimeo. FO: MOR .73/016, Rabat-Rome FAO.

Sarson, M., Salmon P., (1977). 'Elevage, pâturage et données de base pour un amenagement sylvo-pastoral dans la zone no. 2 Tanghaya (Rif Occidental)'. Mimeo. Note technique no. 20, MOR 73/016, Rabat FAO/Min. Agric. Ref. Agric.

Sarson, M., Salmon P., (1978). 'Rôle des arbres et des arbustes fourragers dans 1'aménagement des paturages naturels en Afrique du Nord'. Mimeo. Supporting paper, Item no. 10, 8th World Forestry Congress, Jakarta.

SCET, (1962). 'Zone de mise en valeur de M'sila, étude phytoécologique - Mimeo. Tab., Pl. Atlas.

Schmidt-Burr, H., (1970). 'Composition chimique des fourrages d'Atriplex en Tunisie'. Arch. Laboratorie de Zootechnie. Institut National de la Recherche Agronomique de Tunisie.

Schmidt-Burr, H., (1972). 'Bericht uber die Arbeit in der Abteilung Tierernährung am Nationalen Landwirtschafftichen. Forschungsinstitut in Tunisien. INTRAT, Project FE/302, Tunis.

Theriez, M., (1972). M., (1966). 'Recherches sur la digestibilité des Opuntia fzcus-indica f. inermis en Tunisie'. Doc. Techn., Inst. Nat. Rech. Agron. Tunis.

Wilson, A.D., (1974). 'Nutrition of sheep and cattle in Australian arid areas'. In: Wilson A.D. (ed.) Studies of the Australian arid zone. II Animal production. Perth, CSIRO. Division of Land Resources Management.

Wilson, A.D., (1977). 'The digestibility and voluntary intake of the leaves of trees and shrubs by sheep and goats. Austr. J. Agric. Res., 28, 501-508.

Woodman, H.E., (1942). ' The composition and nutritive value of feeding stuffs'. Min. Agric. Fish. Bull. 124. London.

Ziani, P., (1966). 'Les plantations d'arbres et d'arbustes autour des points d'eau a créer dans le gouvernorat de Medenine'. Note technique no. 4. Inst. de Reboisement de Tunis.