Livestock Research for Rural Development 15 (5) 2003

Citation of this paper

Determination of the quantity of scavengeable feed for
 family poultry on free range 

O A Olukosi and E B Sonaiya

Department of Animal Science, Obafemi Awolowo University 220005, Ile Ife, Nigeria
yinkaolukosi@yahoo.com 


Abstract

 

An eight-week on-farm study was carried out to determine the quantities of scavengeable feed (SF) available on the free range in four villages in south-western Nigeria. The quantity of available SF was determined by direct measurement (SFm) and by calculation (SFc), and these were compared with the predicted SF (SFp). The SF predictors used were the quantity of household leftovers (H, g/household/day), vegetation cover (V, plant stands/m2), number of refuse heaps (R, number/km) and the quantity of crawling insects (I, g/m2/day).

 

The coefficient of multiple determination was high and significant (R2 = 0.99). Inclusion of H as the first factor in the prediction equation increased the precision of the equation the most followed by V and I. However, R is not a good predictor of SF as its inclusion in the equation consistently decreased the precision of prediction. The average quantities of SF determined by the different methods were 20.1, 32.2 and 21.0 g/chicken/day for SFm, SFc and SFp, respectively. There was no difference between the quantities of SFm and SFp, but SFc was significantly different from SFm and SFp.

 

The results indicate that  H, V and I can be used as predictors of the quantity of SF. 

 
Key words: Chicken, feed predictors, free range, scavenging

 

Introduction 

Family poultry serve many families in the developing countries providing them with high quality protein products, income and cultural services. Scavenging chickens utilize certain materials on the free range as feed. Such materials constitute the Scavengeable Feed Resource Base (Roberts 1992). In this paper a shorter abbreviation SF will be used to represent SFRB. The materials that make up the SF include mainly the household leftovers, and other materials like insects, worms (especially earthworms), succulent vegetation, flowers, fruits and crop by-products from cultivation and processing activities (Sonaiya et al 2002a).

 

Roberts (1992) suggested two methods for determination of the quantity of SF. The first method uses the weight of household leftovers and its proportion in the crop content. The second uses the productivity of the scavenging chickens. The first method requires sacrificing the chickens in order to analyze the crop content while the second method requires serial data on life performance. Both methods depend on direct handling of the live birds or their carcasses.

 

To avoid handling live birds or carcasses, it should be possible to use independent parameters to predict the quantity of SF. Sonaiya et al (2002b) listed some possible SF predictors such as the quantity of insects, vegetation cover, refuse heaps, quantity of household leftovers and stage of agricultural activity.  The objective of this study was to compare predicted SF with SF determined by the two methods of Roberts (1992).

 

Materials and Methods 

An eight-week farmer-and-researcher-managed on-farm study was carried out in four villages (Moro, Yakooyo, Ipetumodu 1 and Ipetumodu 2) in the rainforest ecozone of southwest Nigeria. The villages had social amenities like tarred roads, public schools, health care centres, electricity and pipe-borne water. Most of the inhabitants of the villages were farmers cultivating annual crops like maize, cassava, yam, cocoyam and vegetables, as well as perennials such as cocoa, kola and oil-palm. The number of households and the total number of chickens they kept were obtained for each of the villages by a census.

 

The minimum sample size for the study was calculated, based on the variance of the study parameters such as growth rate and egg production, using the formula of Snedecor and Cochran (1967):

 N = 4pq/L²

where:

            N= minimum sample size,
            p = estimate of the population standard deviation,
            q = 100-p,
            L = confidence probability level.

 

The sample size calculated for the study, based on the values of such parameters as egg weight (38 g) and hen-day egg production (12%), was 384. However, 483 birds were actually included at the beginning of the study.

 

Thirty-four free-range flock owners who had at least 10 chickens and who offered some form of shelter and feed supplement were randomly selected across the four villages. Household leftovers were collected from 96 randomly selected households across the four villages at weeks 4 and 8 of the study. The households were provided with plastic waste-baskets for a 24-hour collection. The collection from each household was labelled, its constituents were visually identified, oven-dried and weighed.

 

Forty-two scavenging chickens (14 cocks, 20 growers and 8 hens) were collected across the four villages. The birds were caught at three different periods of the day (9-11 a.m., 1-3 p.m., and 4-6 p.m.) while they were still scavenging and were immediately slaughtered. The crop of each bird was opened up and the crop content was visually identified and manually separated into materials from households and from the environment. These were oven-dried and weighed. The Gross Energy value of the crop content was determined in a ballistic bomb calorimeter (AOAC 1994).

 

The SF (g/chicken/day) based on measurement of household leftovers (SFm) was obtained from the equation of Roberts (1992):

SFRB = [H/P]*[n/(n-x)]

where:

            SFRB = scavengeable feed resource base (kg/family flock),
            H = quantity of household leftover (kg/day),
            P = proportion of H in the crop content,
            n = total number of household in the village,
            x = number of households without chickens.

 

as modified by Sonaiya et al (2002b) to use the total number of chickens instead of number of flocks in the village:

SF = [H/P]*[n/T]

where:

            SF = Scavengeable feed (g/chicken/day),
            H, P and n = as defined above,
            T = total number of chickens in the village.

 

The SF (kg/family flock/year) derived from calculations based on life performance of the birds (SFc) was obtained using Roberts (1992) formula:

SFRB = [ ∑ Ej ] / Es

where:

            SFRB = Scavengeable feed resource base (kg/family flock/year),
            j = average flock size,
            Ej = the ME requirement for the daily maintenance and production of each bird (kcal/bird),
            Es = the ME in the scavenged feed (kcal/kg dry weight).

 

The Ej was calculated from the data on performance and productivity of scavenging chickens obtained over a period of eight weeks using the formula of NRC (1994):

E = W0.75(173-1.95T) + 5.5∆W + 2.07EE        where:

            E = Ej defined above,
            W = average body weight (kg),
            T = ambient temperature (
°C),
           
DW = body weight gain (g/bird/day),
            EE = daily egg mass (g).

 

The Es (kcal/kg) was calculated from the gross energy (GE, kcal/kg) value of the crop content. The ME value was calculated from the GE value using the formula in Larbier and Leclerq (1992): 

ME = 6.913*GE – 18.5*CP – 109.5*CF

where:

            ME = metabolizable energy (kcal/kg),
            GE = gross energy (kcal/kg),
            CP = % crude protein,
            CF = % crude fibre.

 

The SF (g/chicken/day) obtained by prediction used the following parameters: quantity of crawling insects (I, g/m2/day), refuse heaps (R, number/km), vegetation cover (V, plant stands/m2) and household leftovers (H, g/household/day)

 

The number of crawling insects was determined by a modification by Olukosi (2002) of the pitfall trap method of Obeng-Ofori (1994).  The insects trapped were identified by orders using the NRI (1996) chart and then counted. Samples of the insects in the different orders were oven-dried and weighed to determine the quantity of insects trapped.

 

Transect mapping of the vegetation cover and elevation of each village was done using the method of Kirsopp-Reed (1994). Ten random throwings of quandrants were made at different vegetation zones in each of the villages. The number of refuse heaps in each village was counted in the process of transect mapping. The detail of the transect mapping and the modified pitfall trap are described by Sonaiya et al (2002b).

 

The quantities of insets trapped were compared across the villages using one-way analysis of variance. Step-wise regression analyses were done to determine the equation that best predicted the SF. The four predictors were combined in different possible ways to eliminate the predictor/s that did not contribute to the precision of the estimate. The total number of possible combinations of the four predictors was determined using the general algebraic formula for factorial combination as follows:

nCr = n / [r(n-r)]

where:

                nCr = exhaustive total number of possible combinations,
            n = total number of available predictors,
            r = total number of predictors to be combined.

 

Therefore, for the two-predictor multiple regression analysis, six different regression analyses were done using the six different combinations of the predictors. For the three-predictor multiple regression analysis, four different regression analyses were done using the four different combinations of the predictors.

 

The model for the regression analysis is:

YR = ά+ β1X1 + β2X2 + β3X3 + ε

where:

            Yr = mean value of the frequency distribution of Y for specified X1, X2 and X3, respectively;
            ά = overall mean;
            β1, β2, β3  =  partial regression coefficients of Y on X1, X2 and X3, respectively;
            ε = random error.

 

Significance was determined at the 5 % probability level. All statistical analyses were done using SAS (1997) procedures.

 

Results 

In all the villages studied, approximately 55% of the respondents kept chickens, with the average flock size of 16±2. Cocks were the fewest of all the categories of chickens kept in the villages studied (Table 1) 

Table 1. Number of households and family flock composition (means±SD) in the villages

Village

NT

NS

Cock

Hen

Grower

Chick

Flock size

Moro

186

108

3±2

4±3

5±6

6±7

18±15

Yakooyo

172

108

3±3

4±3

5±6

6±8

18±15

Ipetumodu 1

204

102

2±3

3±3

4±5

7±8

16±15

Ipetumodu 2

200

103

1±2

3±2

4±5

5±6

14±13

Average

 

 

2±1

3±1

5±1

6±1

16±2

NT = Total number of households;  NS = Households with scavenging chickens

 Table 2 shows the value of the four predictors of SF in the villages.  

Table 2. Means (±SD) of four predictors of the quantity of scavengeable feed in four villages in south western Nigeria

 

 

Village

Refuse heap, number/km

Crawling insects, g/m2/day

Vegetation cover, plant stands/m2

Household leftovers, g/day

Moro

1.04

7.43±4.0ab

100±29

50

Yakooyo

0.63

9.72±6.6a

109±40

180

Ipetumodu 1

1.64

5.29±3.3b

  98±30

120

Ipetumodu 2

0.95

9.00±3.4a

115±50

110

Average

1.07±0.42

7.86±2.0

106±8

110±50

ab Means in the same column without common letter are different at P<0.05

The pattern of the abundance of R in the villages was related to the number of households in the villages. The same type of insects were trapped in all the villages, and they included insects of the orders Hymenoptera, Orthoptera, Gryllidae, Diptera,, Isoptera, Coleoptera, and Aranaceae. The insects in the order Hymenoptera were the most abundant of the insects trapped (45% of the total), while insects of the order Aranaceae were the fewest in proportion (1.65%). Other metazoans trapped included members of the phyla Mollusca and Annelida. The quantity of insects trapped in Yakooyo appeared to be the highest, but was not different (P>0.05) from the quantity of insects trapped in Ipetumodu 2 and Moro. Ipetumodu 1 had the lowest (P=0.05) quantity of crawling insects trapped.

 

Table 3 shows the result of the step-wise multiple regression analysis using the SF predictors.  

Table 3. Multiple regression analysis using three predictors of the quantity scavengeable feed 

Predictor

b

Sb

tsb

R2

SE

1st combination of predictors

R

34.1

46.5

7.35 (ns)

0.99

1.45

V

-6

1.10

-6.02 (ns)

I

94

13.4

7.08 (ns)

2nd combination of predictors

R

52.6

1.49

35.4 (ns)

0.99

0.26

I

10.8

0.33

32.8 (ns)

H

0.11

3.17

34.5 (ns)

3rd combination of predictors

R

15.22

0.19

78.6*

0.99

0.10

V

0.85

0.005

100*

H

0.12

0.002

81.5*

4th combination of predictors

H

0.13

0.005

-197*

0.99

0.04

V

1.2

0.05

261*

I

-4.4

0.02

220*

b = partial regression coefficient of the predictor; sb = standard error of b
tsb = calculated t-value of sb; R2 = coefficient of multiple determination
SE
  = standard error of the estimate; ns = not significant (P > 0.05)
* = significant t-value (P =
0.05)

 In the 3-predictor equations, including H in the equation markedly increased the precision of the estimate. When H and V were used, the precision increased still more than when H was used alone. The combination of H, V and I gave the most precise estimate. The use of R in the equation decreased the estimate’s precision and hence was not included in the prediction equation. The best prediction equation generated was:

SF = -86 + 0.13H + 1.2V -4.4I

where:

            SF = scavengeable feed (g/chicken/day),
            H = quantity of household leftovers (g/household/day),
            V = vegetation cover (plant stands/m2),
            I = quantity of insects (g/m2/day).

 

The standard error of the estimate was  ±0.04g, and the coefficient of multiple determination was high and significant (R2 = 0.99).

 

Table 4 shows the values for SFm, SFc, and SFp.  

Table 4. Values for the quantities of scavengeable feed (g/chicken/day) determined by various methods in four villages in south-west Nigeria

Village

SFm 

SFc

SFp

Moro

  7.03

28.92

  7.81

Yakooyo

24.50

39.91

25.43

Ipetumodu 1

23.08

31.84

23.92

Ipetumodu 2

25.86

28.22

26.70

Average

20.1±8.93a

32.2±5.36b

21.0±9.00a

ab Means in the same row with different superscripts are significantly different (P=0.05).

The values for SFm and SFp were not significantly different, but SFc was, on the average, 11.2g/chicken/day higher than both SFm and SFp (P = 0.05).
 


Discussion 

 

The census of the households in the villages showed that the flock size (16) is within the range of 15 to 20 chickens per household reported by Sonaiya et al (1993) for this region. The quantity of different scavengeable materials have been measured. In his study of scavenging guinea fowls in Nigeria, Ayeni (1982) found from the crop content that the scavenged materials were made up of 35% grain seeds, 21% insects, 17% cyperus bulbs, 13% fruits, 9% leaves, 2% pebbles, and 3% water ingested with food. Gunaratne et al (1993) found that metazoa made up 8% of the crop content of scavenging chickens in a Sri Lankan community. De Vries (2000) found in his behavioural study of scavenging chickens on the range in Nicaragua that birds spent on average 37% of their feeding time consuming insects. Household leftovers are a very significant contributor to SF and make up between 38 and 72% of the crop content of scavenging chickens in different environments (Gunaratne et al 1993; Sonaiya et al 2002a). It is well known that scavenging chickens peck at fresh succulent swards of grass and scratch refuse heaps for feed materials. Measurement of the predictors should therefore provide an estimate of the SF that is independent of the bird.

 

The similarity in the values of SFm and SFp could be attributed to the similarity in the methods of their determination. The formula for SFm includes both the household and environment components of SF (Sonaiya et al 2002b). Predicted SF measures directly some of the scavengeable materials on the range (insects, household leftovers, vegetation). Therefore, using this method makes it unnecessary to sacrifice chickens in order to analyze their crop content. It may be possible to use prediction equations to determine the quantity of available SF on a range where there are no households, or by implication, no household leftovers (H). This latter value required by the measurement method of Roberts (1992).

 

On the other hand, the SFc determines SF on the basis of the Metabolizable Energy (ME) requirement for performance of the chickens. Indeed, the presence of the variable Ej (which measures ME requirement) makes it apparent that SFc actually determines the SF required for the expected level of production of full-fed, intensively reared birds rather than the available SF.

 

In this study, the quantities of insects (g/day) and household leftovers (g/day) were more directly related to SF (g/day) than the measures of vegetation cover (plant stands/m2) and refuse heaps (number/km). In the case of R, it is the scavengeable materials that are obtainable from the refuse heaps, and not the refuse heaps themselves, that are of interest to the scavenging chickens. The pitfall trap method used in this study is a standard entomological method used mainly for crawling insects. Thus the standard entomological methods for determining the population of flying insects may be used along with the pitfall trap since flying insects are also caught and eaten by chickens on the range.

 

Finally, since the objective of the determination of SF is to know what is available per unit of land, it might be better to express it on the basis of per hectare or per square meter. Methods of evaluating the scavengeable materials on the range per unit area may need to be developed. Eventually, it would be desirable to predict both the quantity and quality of SF on the range.

 


Acknowledgements

 

The authors appreciate the willing cooperation of the 34 individuals in the four villages whose flocks were used for this study.

 


References

 

AOAC  1994 Association of Official Analytical Chemists: Official Methods of Analyses. 14th edition. Arlington. U.S.A.

 

Ayeni J S O 1982 The biology of the helmeted guinea fowl. Ph.D. Thesis, Department of Wildlife and Fisheries, University of Ibadan, Ibadan, Nigeria.

 

De Vries H 2000 Observation on behaviour and feed intake of chickens kept on free range in Muy Muy, Nicaragua. Proceedings XXI World’s Poultry Congress, Montreal, Canada, CD-ROM.

 

Gunaratne S P, Chandrasiri A D N, Hemaltha W A P M and Roberts J A 1993 The feed resource base for scavenging chickens in Sri Lanka. Tropical Animal Health and Production 26: 249-257.

 

Kirsopp-Reed K 1994 A review of PRA methods for livestock research and development. RRA Notes 20:11-36.

 

Larbier M and Leclerq B 1992 Nutrition and feeding of poultry. Nottigham University Press, Leicestershire, U.K.

 

NRC 1994 Nutrient Requirement of Poultry. Sub-Committee on Poultry Nutrition, National Academy Press, Washington D.C.

 

NRI 1996 A guide to insect pest of Nigerian crop. Identification, biology and control. National Resources Institute,  Chatham, U.K.

 

Obeng-Ofori D 1994 The behaviour of 9 stored product beetle at pitfall trap arenas and their capture in millet. Entomological Experimentalis et Applicata 73:39-50.

 

Olukosi O A 2002 A multisystem approach to the determination of the nutritional base for scavenging family poultry. M. Phil Thesis, Obafemi Awolowo University, Ile Ife, Nigeria.

 

Roberts J A 1992 The scavenging feed resource base in assessment of the productivity of scavenging chickens. In: Spradbrow P B (ed) Newcastle disease in village chickens – control with thermostable oral vaccine. ACIAR Proceedings No. 39, Canberra, Australia, pp. 29-32.

 

SAS  1997 SAS Users’ Guide, SAS Inc., North Carolina, USA.

 

Snedecor G W and Cochran W G 1967 Statistical Methods. Iowa State University Press, Ames, Iowa.

 

Sonaiya E B, Laogun E A, Matanmi O, Daniyan O C, Akande B E, Oguntade E A, Omoseibi R O and Olori V E 1993 Health and husbandry aspects of village extensive poultry production in south-western Nigeria. In: Padey V S and Demey F (eds) Proceedings of an International Workshop on Village Poultry Production in Africa, Rabat, Morocco, pp. 34-41.

 

Sonaiya E B, Dazogbo J S and Olukosi O A 2002a Further assessment of scavengeable feed resource base; In: Characteristics and parameters of family poultry production in Africa. Results of a FAO/IAEA Coordinated Research Program, IAEA, Vienna, Austria, pp. 193-200.

 

Sonaiya E B, Olukosi O A, Obi O, Ajuwon K M 2002b Vaccination and scavengeable feed resource assessment for village poultry; Proceedings 3rd Scientific Coordination meeting of the Joint FAO/IAEA Coordinated Research Program on assessment of the effective of vaccination again Newcastle disease and Gumboro disease using immuno-assay-based techniques for increasing farmyard poultry production in Africa. Quatre Bornes, Mauritius, available at: www.iaea.org/programmes/nafa/d3/mtc/sonaiya-doc.pdf



 
Received 16 April 2003; Accepted 31 May 2003

Go to top