Livestock Research for Rural Development 14 (4) 2002

Citation of this paper

Chemical composition and in vitro dry matter digestibility of local landraces of sweet sorghum in Botswana

O R Madibela*,  W S Boitumelo, C Manthe* and I Raditedu

Sebele Research Station, Department of Agricultural Research
P/Bag 0033, Gaborone
, Botswana
*Current address:  Botswana College of Agriculture
P/Bag 0027, Gaborone, Botswana
Corresponding author:


Sweet sorghum is a variant of grain sorghum (Sorghum bicolour (L) Moench, grown in Botswana for edible juicy sweet stem instead of grain. Suitability of sweet sorghum as forage for livestock was investigated by evaluating twelve local landraces for chemical composition and in vitro dry matter digestibility (IVDMD). The samples were collected from a randomised complete block design trial that had three replicates. Samples of whole plant and parts (leaf and, stalk and panicle) were analysed for crude protein, calcium, phosphorus, neutral detergent fibre (NDF), acid detergent fibre (ADF), acid detergent lignin (ADL), neutral detergent bound nitrogen (NDF-N), IVDMD and ash.

The different landraces had similar chemical composition and in vitro dry matter digestibility. Neutral detergent fibre was significantly different among the landraces and was highest in Moitshopari (645 g/kg) and lowest in Botala landrace (550 g/kg). IVDMD averaged 78% and was not significantly different among landraces. Leaf had significantly higher calcium, ADF, NDF, NDF-N, Ash  and phosphorus. Crude protein tended to be different between plant parts. There was a significant interaction between landrace and plant parts for phosphorus only.

The results showed that local sweet sorghum landraces might provide a viable energy resource for ruminants due to high IVDMD. It is necessary to screen these local landraces for biomass production and nutritional evaluation through animal trials.

Key words: Sweet sorghum, forage, chemical composition, digestibility, Botswana


As green forage/fodder is only available for up to a maximum of five months in a year in Botswana, ruminant animals have to depend on cereal straws and other crop residues of poor quality for sustenance. Crude protein (CP) of natural grass during the dry season is as low as 40 g/kg and is only a little higher (70 g /kg) during the wet season (APRU 1975). 

The benefit that may be obtained from the low quality roughages such as cereal stover and natural grass hay is limited by insufficient intake of the roughage for satisfactory animal performance (Ngwa and Tawah 1990). To improve the utilisation of the low quality roughages, the adapted strategy has been to supplement with protein sources such as legume foliages: Lablab (Lablab purpureus) (Boitumelo et al 1993), Sesbania sesban, leucaena (Umunna et al 1995; Bonsi et al 1994), acacia pods (Ndlovu and Sibanda 1996) or agro industrial by-products. However. supplementation with these foliages can only support maintenance and limited live weight gain or production. It is important to provide deficient nutrients to balance the ratios of nutrients absorbed in the small intestine. In many cases, when nitrogen requirements are met through supplementation, energy becomes limiting. Therefore a cheap source of energy will be required to balance the nutrient supply. 

Sorghum (Sorghum bicolour) is one of the main staples for the world’s poorest and most food insecure people (FAO and ICRISAT 1996). In Botswana, alongside grain sorghum, sweet sorghum is also produced. According to Zhu Cuiyun (1998) sweet sorghum is a type of grain sorghum belonging to Graminaceae and it's stem is full of sweet juice. The crop is cultivated widely throughout the world and the stalk is used for producing syrup and as livestock feed (Zhu Cuiyun 1998). In Botswana, sweet sorghum is a marginal crop and is used as sweet reed for sucking the nectar just before the grain reaches the physiological dead ripe stage. Though in recent years the crop has found commercial value for sucking nectar, it is of minor economic importance and hence little attention has been paid to it. However, it was reported by Rose (1982) that sweet sorghum was cultivated as early as the 19th century and Batswana farmers consume it as well as watermelons as a source of food while waiting for grain crops to ripen. 

The grain sorghum absorbs moisture more effectively than other cereal crops such as maize because a sorghum plant has twice the amount of secondary roots on each primary root (Meeske and Basson 1995). According to Black et al (1980) sorghum goes dormant during periods of drought and resumes growth as soon as there is sufficient moisture. Another advantage is that leaf transpiration of sorghum is one half as great as that of maize. It is likely that these attributes are endowed in sweet sorghum as well. Moreover, the local sweet sorghum has been used over years and farmers may have selected varieties adapted to the drought conditions of Botswana. These attributes can be harnessed to develop sweet sorghum for other uses.  

When compared to a popular silage maize variety in China, sweet sorghum variety White Crane produced 20-33% more biomass when grown on reclaimable pasture (Zhu Cuiyun 1998). Ramesh et al (1994) found that the two entries of sweet sorghum, J-Set-3 and Brandies, produced a total (main crop and ratoon) of 15.7 and 15.3 tonnes/ha dry fodder, respectively. 

There are two possible potential ways in which this resource can be used as animal feed. Traditionally in Botswana, when sweet sorghum is harvested, the leaves are stripped from the stalk and are left in the field. These leaves can be used on a cut and carry basis to feed ruminants. Alternatively, the whole plant can be used as hay or to make silage. The results presented here are preliminary data on the chemical composition and dry matter digestibility of local landraces of sweet sorghum in Botswana. 

Materials and methods

The experimental materials were collected from a trial carried out in the 1997/98 planting season at Sebele Research Station in the southern part of Botswana. The station is located at a latitude of 2433" S, longitude 2557"E and at an altitude of 994 masl. Mean rainfall for the area is 500mm but for the 1997/98 planting season it was 367mm. 

The trial was a randomised complete block design with three replications. Twelve landraces were planted on 15/01/97. The plants were established in plots of  10 x 3 m and each landrace was planted in 4 rows at 0.75 m spacing. Samples taken randomly from each replicate were bulked for analysis at dough stage on 11/05/98. This was the start of winter. Sample materials were separated into leaves, stalk and panicle or used as the whole plant. Whole plant and stalk samples were air-dried for 3 days and then transferred to an air blown oven for further drying at 50C for 3 days. Leaves were oven-dried at 50C for 3 days. Contents of DM, organic matter (OM) neutral detergent fibre (NDF), acid detergent fibre (ADF), and acid detergent lignin (ADL), NDF-nitrogen, nitrogen (Kjeldahl method) and minerals (wet digestion) were determined following the methods of AOAC (1996). Calcium was determined with an atomic absorption spectrophotometer and phosphorous with an ultra-violet spectrophotometer. In vitro dry matter digestibility (IVDMD) was estimated according to Tilley and Terry (1963), by incubating in a thermostatically controlled circulating water bath. Rumen liquor was donated by a Tswana steer fed a mixture of Cenchrus ciliaris and Lablab purpureus at a ratio of 1:1 at a rate of 8 kg/day. The diet contained (g/kg DM) 76.8 crude protein (CP), 7.8 calcium, 1.2 phosphorous, 777 NDF, 547 ADF and 904 OM. 

Statistical analysis

The General Linear Model (GLM) procedure (SAS 1990) was used to test the effects of landrace and plant parts on chemical composition and IVDMD. Differences between specific landraces and between plants parts were tested for significance using Duncan’s Multiple Range Test. Interactions of landraces and plants parts were also tested. Correlation between chemical attributes was also tested. Means are reported as least square means. 


NDF content was the only chemical constituent showing significant differences among the landraces (P<0.05). NDF was highest for landrace "Moitshopari"  and lowest for landrace "Botala". Differences in IVDMD among landraces were not significant, averaging 78.0% (Table 1).. 

Table 1.  Chemical composition and in vitro dry matter digestibility of twelve local sweet sorghum landraces#

Khibidu 74.0 68.2 600.2 294.1 30.0 8.0 3.0 1.4 77.5
Maidiwa 77.2 69.6 560.0 267.7 31.6 8.2 3.3 1.4 78.5
Makoro 91.5 70.7 606.1 289.4 33.3 8.9 3.3 1.5 72.6
Mmamaidi 74.2 72.3 618.5 296.8 28.8 9.3 2.7 1.5 76.8
Ostrich 69.4 70.3 570.5 277.5 28.6 9.1 2.4 1.2 78.8
Nata Z 78.3 66.0 583.9 284.3 32.5 8.7 3.4 1.5 77.8
Sebethwane 85.7 77.1 594.2 285.7 31.4 8.8 3.0 1.5 77.7
Mmankhibidu 89.1 70.8 608.9 294.4 34.0 7.2 3.6 1.6 77.1
Moitshopari 83.7 62.4 644.9 324.4 29.0 7.1 3.7 1.1 73.1
Phuthamogwang 2 75.7 67.6 586.9 277.4 27.5 8.3 3.4 1.5 80.2
Botala 82.9 75.3 547.9 260.0 39.8 9.2 3.2 1.5 82.2
Kedikilwe 78.7 82.3 565.7 268.9 25.5 10.0 3.5 1.6 83.2
Mean 80.0 71.0 589.1 285.1 31.0 8.6 3.2 1.5 78.0
SEM 5.9 5.6 17.8 13.3 6.4 1.0 0.4 0.2 3.2
Significance NS NS * NS NS NS NS NS NS
NS = P>0.05  *=P<0.05
CPa  = Crude protein, Ca = calcium, P =phosphorous, ADF = acid detergent fibre, ADL = acid detergent lignin, NDF = neutral detergent fibre
NDF-N = neutral detergent fibre-nitrogen, DMD = in vitro dry matter digestibility.
# Units are expressed as g/kg DM except for DMD, which is in % DM

The leaf part had significantly higher calcium, ADF, NDF, NDF-N and ash (P<0.001) than other parts. However, the stalk and panicle, and the whole plant had significantly higher (P<0.05) phosphorous and IVDMD values than leaves (Table 2). CP tended to be different between plant parts (P=0.053). 

Table 2. Chemical composition and in vitro dry matter digestibility of the leaf, stalk and panicle, and whole plant fractions of sweet sorghum1
Leaf 115.5 76.8 691.7 344.6 31.1 10.3 5.5 1.3 74.0
Stalk 60.2 67.4 536.6 255.3 35.5 8.2 2.0 1.6 80.7
Whole plant 64.4 68.9 543.6 255.2 26.4 7.1 2.2 1.5 79.2
Mean 80.0 71.0 589.1 285.1 31.0 8.6 3.2 1.5 78.0
SEM 3.0 2.8 9.0 6.7 3.2 0.5 0.2 0.1 1.6
Significance *** NS *** *** NS *** *** * *
NS = P>0.05    * = P<0.05   ** = P<0.01   *** = P<0.001
CPa  = Crude protein, Ca calcium, P =phosphorous, ADF = acid detergent fibre, ADL = acid detergent lignin,
NDF = neutral detergent fibre
NDF-N = neutral detergent fibre-nitrogen, DMD = in vitro dry matter digestibility.
1Units are expressed as g/kg DM except for DMD, which is in % DM  

There was a significant (P<0.05) interaction between landraces and plant parts for phosphorus only. There tended to be an interaction (P=0.083) between landrace and plant parts for CP. There was a significant correlation (r=0.68, P<0.001) between CP and NDF-N but a weak (r=0.27, P<0.05) correlation between CP and NDF. The IVDMD was negatively correlated to both NDF and ADF (r= -0.50, P<0.001). 


In the present study the CP levels of the forage harvested at 16 weeks were similar between landraces but were higher than that of wild sorghum harvested 4 weeks earlier by Fleischer and Timpong (1996). These authors found that at 8 weeks CP was higher which suggests that it may be beneficial if harvesting commenced earlier for landraces reported in the present study. The mean CP of 71.0 g/kg in the present study is lower those of grain sorghum stovers tested by Youngquist et al (1990) in Botswana but higher than the mean of 62.0 g/kg of grain sorghum residue reported by Snyman and Joubert (1995) in South Africa. It was also well above the 44g/kg level in natural grass harvested at full maturity (APRU 1975). It is in the minimum level permissible for adequate  feed intake and rumen function considering the high IVDMD (78.0%) and that as a much as 0.9% of the nitrogen was in the NDF fraction. Otherwise the high proportion of NDF-N would probably indicate that an important fraction of the undegradable protein would be also unavailable to the host animal. Adewakun et al (1989) and Felix and Funso (1994) reported low CP in silages made of sweet sorghum using varieties Brandes and Theis, compared to maize or grain sorghum silage or fescue hay. However, Adewakun et al (1989) found lower urinary N excretion and greater N retention for steers fed sweet sorghum silages and concluded that protein from sweet sorghum silage may be more efficiently utilised than that from maize silage. 

Minerals, calcium (Ca) and phosphorus (P), which form a major mineral supplementation intervention by Botswana farmers in the form of licks containing dicalcium phosphate were the same among the landraces. Calcium levels were however lower than 4.6g/kg reported by Youngquist et al (1990) for grain sorghum stover in Botswana. APRU (1980) reported levels of P from pasture clippings over several years that rarely reached 1.0g/kg DM. In the present study, the P levels were higher than the 1.2 g/kg reported by Youngquist et al (1990) and Snyman and Joubert (1995). The leaf fraction had higher Ca levels, four times higher than those of P, which could lead to an imbalance in the diets of ruminants fed on sweet sorghum alone. However, the whole plant had Ca and P levels that had a ratio of 1.5:1 and is within the range 1:1 to 2:1 recommended for ruminants (McDonald et al 1988). 

Though ADF was similar among landraces, NDF was different. The fibre components and lignin were lower than values reported for wild sorghum (Fleischer and Timpong 1996), grain sorghum stover (Snyman and Joubert 1995) and sweet sorghum silages (Felix and Funso 1994). The limitation imposed by high fibre content is the reduction in dry matter digestibility leading to insufficient supply of energy. As a consequence IVDMD was negatively correlated to NDF and ADF. In the present study, IVDMD levels of all landraces were far higher than 45% reported by Youngquist et al (1990) to be the acceptable level to maintain weight of cattle in the tropics. The values observed in the present study were also higher than the 73.3 and 72.5% reported by Felix and Funso (1994) with sheep. The fact that the stalk/panicle (stem) has a high level of fermentable sucrose (Zhu Cuiyun 1998), in addition to the grain in the panicle may be the principle reason for high digestibility of the stalk/panicle and the whole plant. The leaf fraction had higher fibre components, which is the reason for their lower IVDMD. Vadiveloo (2000) found low IVDMD (cellulase neutral detergent method) of the leaf compared to the stem of rice straw. In the present study, the lower IVDMD of the leaf fraction may also be due to the high level of total ash observed, which included insoluble ash. There is need to explore other ways of utilising the valuable attributes of sweet sorghum for livestock feeding.  This is an important proposition for Botswana conditions where grain sorghum stover is used for ruminant feed but has a limitation of low nutritive value. 


The results of the present study show that sweet sorghum landraces found locally might represent an acceptable source of energy for ruminants, although their fibre level could be limiting for optimum fermentation in vivo. The fact that the leaf fraction had higher levels of fibre components but lower IVDMD may mean it would not be advisable to feed them as sole feed. Though chemical attributes of tested landraces were similar, their biomass production is not known. Visual appraisal by the authors indicated that some landraces have potential to produce reasonable biomass to warrant use as animal feed. It is therefore recommended that experiments be designed to look at suitable stage of harvesting together with biomass production and nutritional attributes. 


The authors would like to acknowledge the technical input of the Plant and Feed Laboratory staff. The study was financed by Botswana’s Ministry of Agriculture.


Adewakun L O, Famuyiwa A O, Felix A and Omole T A 1989 Growth performance, feed intake and nutrient digestibility by beef calves fed sweet sorghum silage, corn silage and fescue hay. Journal of Animal Science. 67:1341-1349

AOAC 1996 Official Methods of Analysis, Association of Official Analytical Chemists, Washington DC.

APRU 1975 Animal Production Research Unit (APRU, 1975). A five year programme of integrated beef cattle and range research in Botswana 1970-1975. Part I, Technical Results. Government Printers, Gaborone.

APRU 1980 Handbook on Beef Production and Range management in Botswana. Government Printers, Gaborone.

Black J R, Ely L O, McCullough M E and Sudweeks E M 1980 Effect of stage of maturity and silage additives upon yield of gross and digestible enegy in sorghum silage. Journal of Animal Science. 50:617-624

Boitumelo W S, Madibela O R and Mahabile W 1993 Potential of crop residues for small scale milk production: dairy research experience. In: G. Tacheba, W. Mahabile and J. Dailey (Editors). The potential for increasing livestock productivity in Botswana. Proceedings of the livestock production workshop held in Gaborone, Botswana, 14-17 October, 1991.

Bonsi M C K, Osuji P O, Nsahlai I V and Tuah A K 1994 Graded levels of Sesbania sesban and Leucaena leucacephala as supplements to teff straw given to Ethiopian Menz sheep. Animal Production. 59:235-244

Felix A and Funso A O 1994 Digestibility and nitrogen balance in lambs fed grain sorghum silage, sweet sorghum silage or fescue hay. Small Ruminant Research. 14:33-38.

Fleischer J E and Timpong J 1996 Studies into the dry matter yield and nutritive quality of ‘wild sorghum’ (Sorghum arundinaceum) and phasey bean (Macroptilium lathryoides). The African Feed Resources Network Newsletter. 6:1-4

FAO and ICRISAT (Food and Agricultural Organisation and International Crops Research Institute for Semi-Arid Tropics) 1996 The world sorghum and millet economics; facts, trends and onlook. FAO and ICRISAT, 5-25pp

Meeske R and Basson H M 1995 Research note; maize and forage sorghum as silage crops under drought conditions. African Journal of Range and Forage Science. 12:133-134

Mcdonald P, Edwards R A and Greenhalgh J F D 1988 Animal Nutrition. Longman Scientific and Technical, Essex. pp 389-391

Ndlovu L R and Sibanda L M 1996 Potential of dolichus lablab (Lablab purpureus) and Acacia tortolis pods in smallholder goat kid feeding system in semi-arid areas of Southern Africa. Small Ruminant Research 21:273-276

Ngwa A T and Tawah C L 1990 Effect of legume crop residues and concentrate supplementation on voluntary intake and performance of Kirdi sheep fed a basal diet of rice. In: Stares, J. E. S., Said, E. N., Kategile, J. A., (eds). The complementary of feed resources for animal production in Africa. Proceedings of the joint feed resources networks workshop held in Gaborone, Botswana, 4-8 March 1991.

Ramesh B, Gumaste S K, Jayanna M, Patil T C and Baba R 1994 Studies on fooder yielding ability of sweet sorghum. Karnataka Journal of Agricultural Science. 7:391-394

Rose E 1982 Second thoughts on the history of Botswana arable agriculture. Botswana Notes and Records. 14:30-33

SAS Institute Inc. 1990 SAS Users’s Guide. Statistics. SAS Institute INC Version 6. Cary, NC.

Snyman L D and Joubert H W 1995 Chemical composition and in vitro dry matter digestibility of various utilization forms of grain sorghum residues. African Journal of Range and Forage Science. 12:166-120

Tilley J M A and Terry R A 1963 A two stage technique for the in vitro digestion of forages. Journal of British Grassland Society 18:104-111

Umunna N N, Osuji P O, Nsahlai I V, Khalili H and Mohamed-Saleem M A 1995 Effect of supplementing oat hay with lablab, sebania, tagasaste or wheat middlings on voluntary intake, N utilization and weight gain of Ethiopian Menz sheep. Small Ruminant Research. 18:113-120

Youngquist J B, Carters D C and Clegg M D 1990 Grain and forage yield and stover quality of sorghum and millet in low rainfall environment. Experimental Agriculture. 26:279-286

Vadiveloo J 2000 Nutritional properties of leaf and stem of rice straw. Animal Feed Science and Technology. 87:57-65

Zhu Cuiyun 1998 Review and perspective on sweet sorghum breeding in China. International Sorghum and Millets Newsletter. No 39. pp 70-71

Received 18 May 2002

Go to top