Milk production performance of Zebu, Holstein Friesian and their crosses in Ethiopia

Million Tadesse and Tadelle Dessie ^*

Debre Zeit Agriculture Research Centre, P.O.Box 32,  Debre Zeit, Ethiopia

Humboldt University of Berlin, Animal Breeding for Tropics and sub-tropics
Philippstr. 13, Haus 9,  10115 Berlin, Germany
sarmeder@hotmail.com

Abstract

In this study, the milk production performance data from 1973 to 1998 of eight breed groups, at Debre Zeit Agricultural Research Centre dairy herd (Ethiopia) was analysed. The aim was to assess variations in different breed groups and define the factors that influence dairy performance traits. The General Linear Model (GLM) of Statistical Analysis Systems (SAS) was used to quantify both breed and environmental effects, including period (year-group), parity and seasons of calving on milk production traits.

The results showed that with the exception of season of calving, all the other factors exerted significant influences on total lactation milk, annual milk and average daily milk yields. The highest estimated least square means of milk yield  (3083 kg for total milk and 2678 kg for annual milk yield) were obtained from Holstein Friesian cows followed by 3/4 Friesian crosses and Barca breed produced the lowest for both total milk (672 kg) and annual milk yield (673kg). On the base of annual milk yield Holstein Friesian cows remain top followed by ½ Holstein Friesian of Barca crosses and Barca cows produced the lowest annual milk yield. Both breed and period (year-groups) have significant effects on lactation length; however the differences between parity and season of calving were not significant. Lactation length ranged from 279 days for Barca to 362 days for Holstein Friesian cows. Calving interval was found to be significantly influenced by breed, parity and period. The minimum and maximum calving intervals were 400 and 498 days for 1/2 Holstein Friesian x 1/2 Barca and 7/8 Holstein Friesian x 1/8 Barca, respectively. Breeding efficiency was highest for F₁ crosses and lowest for high grade and Holstein Friesian cows.

Under the prevailing feeding and management conditions, cows with 50 to 75 % Friesian inheritance would be recommended provided that sound breeding programmes are in place to maintain the recommended blood level.

Keywords: Crossbreeding, Ethiopia, Milk production, Zebu

Introduction

The cattle indigenous to most tropical countries belong to the species Bos indicus. This species is well adapted to tropical environments. It possesses a high degree of heat tolerance, is resistant to tick borne and to other diseases occurring in the tropics, and has a low maintenance requirement. However, its potential for milk production is low. On the other hand, Bos taurus (European type) is the predominantly specialised dairy breed of the temperate countries. These breeds have high milk yield potentials but lack heat tolerance and disease resistance.

One way of improving tropical cattle regarding milk production is through crossbreeding with Bos taurus dairy breeds. This has been widely used in order to combine the high milk yield potential of exotic breeds with the adaptability of the local ones. The first crossbred generation (F₁), usually from native females mated with exotic males, has been a success in most cases. The F₁ crosses can produce up to three times more milk, and have longer lactation and shorter calving intervals than the local breeds (Kiwuwa et al 1983). However, back crossing to the European breeds gave rather disappointing results; i.e. milk yield increased only slightly or even declined, and fertility deteriorated. This is in addition to the lack of adaptation to tropical conditions (Syrstad 1989).

In Ethiopia, there are eight cattle breeds and most of them belong to the Zebu types with some Sanga and Sanga-Zebu intermediate (Alberro and Haile-Mariam 1982); however, their potential for milk production is low (IAR 1976).

Two of these local breeds, Barca and Boran are not only the main sources of milk but also meat and draft power in Northern and Southern Ethiopia, respectively. Recognising their importance the then Alemaya college of Agriculture, Debre Zeit Agricultural Research Centre launched a breeding program with Local and Holstein Friesian to upgrade their milk production performances and to test the performance of pure (Barca and Friesian) and upgraded animals. The objective of this paper is, thus, to examine the milk production performance of these pure and cross animals and the factors that are influencing the dairy performance.

Materials and Methods

The study area

The data used in this study came from the crossbreeding program of the Debre Zeit Agricultural Research Centre (DZARC), in the central highlands of Ethiopia. Debre Zeit is located at 8°44’latitude and 39°02’ longitude with an elevation of 1900 m.a.s.l.  The topography is generally flat with many small crater hills and mountains. The dominant soils, which occur in flat and gentle slopes, are seasonally waterlogged black clay soils or Vertisols with high potential for cereals, and on some gentle and steeper slopes light soils or Andosols dominate. The climate is characterised to be bimodal with two rainy seasons in a year. The short rainy season called belg occurs between March and May and the main rainy season is during July to September. The average annual rainfall during the last 35 years is 825.4mm, the average monthly temperature was 17.98° and the average monthly relative humidity was 52.4%.  Crop production is the dominant activity. Livestock also plays an important role in the farming system.  

Establishment of the herd, breeding programme and breed groups 

The herd was established in 1971 with eight pregnant Holstein Friesian (F) and seven F₁ Holstein Friesian * Boran (FBo) cross-heifers obtained from Alemaya College of Agriculture (now Alemaya University of Agriculture) in Ethiopia. In 1972, 28 Barca (Ba) heifers were bought from Eritrea and were mated to Ba bulls to generate station-born heifers. Subsequently, both parental cows and their first female offspring were assigned to F bulls to generate FBa, FFBa and FFFBa crosses. Pregnant FBo crossbred heifers (F1) bought from Abernossa ranch in the Ethiopian Rift Valley, were also included in the same year. The FBo crosses were further upgraded to FFBo and FFFBo crosses. Breed groups and number of cows included in the analysis are indicated in Table 1.

Management and feeding of the animals

All animals were assigned individual ear tag numbers. At each calving, identity of the dam and sir, breed, sex, the date and identity of the calf were recorded. The feeding practice was designed in such a way that it gives continuous growth, with feed being offered in-group, regardless of breed. Hay and maize silage constituted the major proportion of the feed supply. Whenever there was a short supply of hay, tef (Erogrostis tef) straw was substituted. At times, green alfalfa and oat hay were also fed. In many instances however, there were feeding irregularities over the years. Milking cows were supplemented with concentrate composed of wheat by-products and Noug seed cake (Guizotia abyssinica) while they were milked. The amount depended on the volume of milk from each cow. In all cases cows were mated naturally and were managed in a loose housing system. Cows were hand milked twice a day. Animals on the farm were regularly vaccinated against common infectious diseases. Regular preventive treatments were administered against prevalent endo- and ecto- parasites.

Data recording, preparation and statistical analysis

Data (623 records) collected from 1973 to 1998 on milk production traits of Barca, Friesian and their Barca*Friesian and Boran* Friesian cows were used for the analysis. All data with the exception of few records were included. After preliminary analysis, a total of 15 records on total milk yield were excluded due to short (less than 150 days) and long (greater than 600 days) lactation. In the first analysis, 26 years of calving (from 1973 to 1998) were used in the model to correct the effect of breed group for year of calving. The results obtained were unreal, and difficult to interpret (for instance the milk production of Barca breed was underestimated). To minimise the gap between the local breed and its crossbreds these calving years were grouped into six periods (year-groups) (Table 3) based on the changes in rainfall distribution and management levels. This grouping technique is adopted with slight modification from that used by Mekonnen and Goshu (1987). In order to see the effects of season of calving, months of the year were classified into four seasons; short rains (March-May), long rains (June-August), post-rainy seasons (September and October), and dry season (November- February).

 

The maximum lactation in the original data was 9 (lactation 1 to 9). However, when 9 lactations were considered in the model, the estimated least square means for lactation numbers 6 and greater than 6 were almost similar. Therefore, we decided to make lactation class 6 and greater than 6 as one group in the final analysis. Calving interval (days) for each lactation was computed as the difference between the start of lactation and next calving.

 

The data were analysed using the General Linear Model (GLM) of SAS (1987). The following mathematical model was applied to analyse all traits:

 

y_ijklm = M + L_i + S_j +P_k + B_l +e_ijklm

 

Where:

y_ijklm = Lactation milk yield, annual milk yield, average daily milk yield, lactation length, calving interval and Breeding efficiency of an individual cow with lactation i, in season j, year groups k of breed group l

M = overall mean        

L_i = the effect due to the i^th lactation number (i = 1...6 )

S_j = the effect due to j^th season of calving (1, 2, 3 and 4).

P_k = the effect due to the k^th year group of calving (k = 1...6)

B_l = the effect due to the l^th breed group (l = 1...8)

e_ijklm = random error effect.

 

Breeding efficiency was computed according to Kiwuwa et al (1983). The estimated coefficients were expressed as percentages.

 

BE = {(N-1) 390 + 960} / (age at each calving in days)

Where:

 

BE = breeding efficiency

N-1 = the number of calving intervals with N calving

390 = is the upper limit of desirable calving intervals (days)

960 = is the upper limit of age at first calving (days)

 

Annual milk yield was calculated as = (Total milk (kg)/calving interval in days)*365 days

 

 

Results and discussion

 

Lactation and average daily milk yield

 

Unadjusted means for milk production traits and breeding efficiency are presented in Table 1.  The estimated Least Square means are in Table 2. The fact that annual milk yield for 100% Barca and for 50% Holstein Friesian and 50% Boran is higher than the total milk yield is due to the difference in the date set used for the two traits. For total lactation milk yield, all lactations were used during the calculation; however, for annual milk yield some lactations do not appear in the model due to lack of an associated calving interval. In a similar way some cows, which had a calving interval without having a corresponding milk yield,were used in the model for calculating calving interval. The lactation yields that were discarded were those with incomplete data, or because of too short or too long lactation length.

Table 1. Unadjusted means for total milk, annual milk yield, average daily milk yield (Milk /ll), calving interval (CI) and breeding efficiency (BE) of Barca, Holstein Friesian, Barca* Holstein Friesian and Boran* Holstein Friesian crosses.
Breed groups	Total milk, kg		Annual milk, kg		Milk/LL, Kg/day		CI, days		BE, %
	N*	Mean	N*	Mean	N*	Mean	N*	Mean	N*	Mean
Barca	35	869	21	1099	32	4.46	43	397	---	---
1/2Barca* 1/2Friesian	109	2055	93	1903	103	6.66	91	415	91	92.45
1/2Boran* 1/2Friesian	87	1740	54	1752	87	5.93	59	440	59	98.99
1/4Barka* 3/4Friesian	87	2214	74	1797	79	6.17	77	474	77	86.55
1/4Boran* 3/4Friesian	129	2044	102	1689	126	5.77	107	471	107	87.08
1/8Barca* 7/8Friesian	36	2381	15	1511	24	5.84	16	512	16	85.54
1/8Boran* 7/8Friesian	35	1902	28	1420	23	5.55	31	493	31	81.21
Holstein Friesian	90	3028	75	2611	81	9.99	82	460	82	82.61
* Number of animals

Table 2 Least Square Means (± SE) for total milk yield, annual milk yield, average daily milk yield (milk/LL), lactation length (LL) and calving interval (CI) of Barca, Holstein Friesian, Barca* Holstein Friesian and Boran* Holstein Friesian crosses.
Breed group	Total milk, kg	Annual milk, kg	Milk/LL, Kg/day	LL, days	Cl, days
Barca	672±196e	674±224e	2.98 ± 0.69e	279± 24c	401± 24cde
1/2Barca* 1/2Friesian	2316± 98bc	2042±106b	7.21 ± 0.26b	326 ± 11bc	400 ±14d
1/2Boran* 1/2Friesian	2088± 118cd	1887±136cd	6.36± 0.30cd	328±  13 b	426±19bd
1/4Barca* 3/4Friesian	2373±105b	1953±111bc	7.15 ± 0.28b	360 ± 12 a	448± 16ab
1/4Boran* 3/4Friesian	2336±96bc	1975±106bc	6.92 ±0.25bc	358 ± 11 a	436±15bce
1/8Barca* 7/8Friesian	2189±183bcd	1558±239bc	6.28±0.52bcd	351±  22ab	498 ± 30a
1/8Boran* 7/8Friesian	1915±163d	1501±173bcd	5.98±0.50c	341±  20 ab	464 ±24ab
Holstein Friesian	3183 ± 111a	2679±120a	9.43 ± 0.39a	362±  13 a	458± 16ab
Mean ± SE	2134± 49	1784±66	6.54 ± 0.15	338±  6	442±8
abcd          Means within a column followed by different superscripts are significantly different *= P<0.05;  **= P<0.01; ***= P<0.001

 

Daily milk yield was lowest for purebred Barca cows and highest for purebred Holstein Friesian. Of the crosses of Holstein Friesian with Barca, the daily milk yields were higher for 1/2 and 3/4 Holstein Friesian than for 7/8 Holstein Friesian (Figure 1), but were less than for the purebred Holstein Friesian. These results for the comparison of F1 with purebred Holstein Friesian are at variance with those reported by Madalena (1993) from an extensive study in Brazil, where F₁ crosses of Holstein*Zebu (Guzerat) produced as well as purebred Holstein, at a high management level, and were better than purebred Holstein at moderate levels of management. The decreases in milk yield of high grade (above 3/4 Holstein Friesian) compared with the F1 might be related to the break down of genes due to epistatic effect; however, since the number of animals available for this group were small, the results needs to be interpreted with caution.

 

Figure 1: Mean daily milk yields per day of calving interval of Barca and Holstein Friesian cows and their crosses

 

The F1 Barca cows had a higher daily milk yield than F1 Boron cows, and this superiority tended to be repeated in 3/4 and  7/8 crosses with Holstein Friesian (Figure 2). These findings are supported by data from Holleta Research Station, in Ethiopia (Beyene 1992), where lactation yields for purebred Barca (675 kg) were superior to those for Boran (494 kg). However, when milk yield was corrected for calving interval (Figure 3), there were no differences between crosses of Barca and Boron, reflecting the superior breeding efficiency of the Boron crosses (Figure 4).

 

Figure 2: Mean daily milk yields of crosses between Holstein Friesian and Barca and Boron cows	Figure 3: Mean milk yields per day of calving interval (CI) of crosses between Holstein Friesian and Barca and Boron cows

 

In his review of the performance of crossbreeds in the tropics, McDowell (1985) concluded that the 3/4 crosses with exotic breeds were superior in milk production compared to F₁ crosses. This is later review is in agreement with the findings of this study. Another work also revealed that cows with exotic blood levels higher than 50% were found to be more productive even under village conditions in India (Mangukar et al 1984). At Arsi livestock breeding station (Ethiopia), the performance level of cows with three-quarter of their inheritance from exotic breeds was higher than that of half-breeds (Kiwuwa et al 1983). However, the above findings and the results of this study are at variance with those reported by Madalena (1993) from an extensive study in Brazil, where F₁ crosses of Holstein*Zebu (Guzerat) produced as well as purebred Holstein, at high management level, and better than purebred Holstein at moderate level of management. Mason and Buvanendran (1982) and Hirooka and Bhutyan (1995) reported higher milk yield by exotic breeds in the tropics when animals were well fed and managed.

 

Total milk, annual milk and average daily milk yield were significantly affected (p< 0.05) by parity (Table 3).  

Table 3. Estimated least square means (± SE) for parity, period (year group) and season of calving of Barca, Holstein Friesian, Barca* Holstein Friesian and Boran* Holstein Friesian crosses.
		N*	Total milk,  kg		Annual milk,  kg	Milk/ll, kg/day	LL, days	CI, days
Parity
1		169		1970±85cde	1511±98b	5.94± 0.24c	364±10a	460± 13a
2		144		2099±85ae	1710±98a	6.43± 0.24b	353±10ab	468 ±13a
3		102		2314±96a	1898±104a	7.10± 0.25a	343±11ac	437±15ab
4		86		2210±105ab	1811±120a	6.64± 0.28ab	333±12bc	458 ±16ab
5		54		2191±125ac	1820±131a	6.90± 0.33ab	317± 14c	415 ± 19b
6+		59		2023±128bcde	1944±161a	6.74± 0.36ab	316± 16c	410 ± 22b
Period
1973-1978		86		2381± 147a	2326±153a	8.73±  0.38a	301± 16c	420± 21cd
1979-1982		189		1730± 95c	1629±103b	6.26±  0.25bc	294±  11c	431 ±14bd
1983-1986		157		1685± 83c	1388±93c	5.07±  0.23d	343± 10b	496± 13a
1987-1990		98		2048 ±100b	1764101b	5.84±  0.28b	353±  12b	453± 14bcd
1991-1994		58		2555 ±131a	1958±140b	6.59±  0.34b	396±  15a	467±  19ab
1995-1998		25		2407± 13ab	1639±330b	6.75±  0.65bc	339±  27bc	382 ± 42cd
Season  of calving
Mar-May	135			2163± 83NS	1852±98NS	6.88±  0.26ab	332±  7.9NS	430±  13b
Jun-Aug	145			2093± 81NS	1734±97NS	6.13±  0.25c	345±  7.5NS	463± 13a
Sept-Nov	105			2122± 94NS	1775±102NS	6.55±  0.35abc	343± 8.2NS	429± 14b
Dec-Feb	225			2159± 68NS	1774±83NS	6.60±  0.21ab	331±  6.5NS	445± 11ab
* Number of animals abcdMeans within a column followed by different superscripts are different at P<0.05

In all cases estimated least square means increased up to third parity, after which total lactation and average daily milk yield declined consistently. First and second lactation milk yields were significantly lower (p< 0.05) and third lactation yields were significantly higher (p < 0.05), than the other lactation yields. Cows that were in the third parity produced 21.5% more total milk than cows in the first parity. On the base of annual milk yield, milk yield increased from parity one to parity three and slightly decreased from parity three to parity four and decreased then after. Cows with parity one still produced significantly lowest annual milk yield than cows with other parities the difference between other parities were not significant for annual milk yield.

 

Mackinnon et al (1996) also reported a decrease in milk yield and lactation lengths after the third parity on crosses of Ayrshire, Brown Swiss and Sahiwal in Kenya. However, the Fogera-Friesian crosses at Gonder Breeding Station (Ethiopia), the maximum lactation yield was attained at fourth and fifth parity and declined then after (Goshu and Mekonnen 1997) Martinez (1988) reported that age difference (i.e. difference of lactation number) is one of the important non-genetic sources of variation in milk yield.

 

In the present analysis, period-effects (year-groups) were highly significant (p< 0.0001) for total lactation, annual and average daily milk yield. Although there was no consistence trend, total lactation milk, annual milk and milk/LL declined between 1973-1986 (period one to three) and increased from 1987 to 1994 and decreased then after. This variation in milk yield over year group of calving might be related with changes in climatic and other management conditions.

 

Season of calving had not significantly (p> 0.05) influenced almost all traits. Similar results were reported on crosses of local and Holstein Friesian cows on milk yield (Hirooka and Bhuttyan 1995). This suggests that even in the tropics, the influence of climatic conditions may be negligible under optimal feeding and management conditions. Although season of calving had significant influence on lactation yield, cows calving during the short rainy seasons (March-May) produced the highest total lactation, annual and average daily milk yields (Table 3)

Lactation length 

Lactation length increased with proportion of Holstein Friesian inheritance from 279 (Barca) to 362 days (pure Friesian) (Table 2). The F₁ crosses had the shortest lactation length compared to other crosses. Lactation length obtained for Barca breed in this study was longer than reports for Boran (155 days) and almost similar to Horro (285 days) cattle in Ethiopia (Beyene1992). However, Kiwuwa et al (1983), reported lactation length of 272 and 303 days for Arsi and Zebu breeds in Ethiopia, respectively. Lactation length decreased slightly with increasing lactation number. The first lactation length was the highest compared to the subsequent lactation lengths (Table 3). Cows calved during the latter periods had significantly (p<0.05) longer lactation length. Season of calving did not show differences on lactation length. 

Calving interval 

Calving interval was significantly (p<0.05) affected by breed group, parity and period of calving (Tables 2 and 3). However, season of calving did not show significant (p>0.05) effect on calving interval. The high grade and Holstein Friesian cows had significantly (p<0.05) longer calving intervals while F₁ and the Barca breed had lower calving intervals. The mean calving interval for Barca breed was 401 days which is lower than 474 days of calving interval for Horro breed in Ethiopia (Goshu and Mekonnen 1997). Kiwuwa et al (1983) and Negussie et al (1998), respectively, reported 439 and 410 days of calving intervals for Arsi breed at Arsi Livestock Breeding Farm. Except half crosses Boran crosses had slightly shorter calving interval than Barca crosses (Table 2). The first and second calving intervals were significantly (p<0.05) higher than the subsequent calving intervals. The sixth and above calving intervals were lower than the elapsed calving intervals. Although there was no clear trained, calving intervals were increased from period one to three (1973 to 1986) and declined then after. The highest calving interval in period three might have been related to the decline in the feed supply and management level encountered in the research centre.

Breeding efficiency

Means of breeding efficiency are given in Table 1. BE were highest for 50% Friesian cross and lowest for 7/8 Friesian x 1/8 Boran cross and pure Friesian cows. The other crossbred animals with different Holstein Friesian inheritance shared intermediate and similar values of BE. These would indicate that crossbreed animals with 50% to 75% Holstein Friesian inheritance gives in better efficiency with respect to reproductive performance, and the Holstein Friesian tended to have rather lower efficiency than crossbred animals.

 

References 

Alberro M and Haile-Mariam S 1982 The Indigenous Cattle of Ethiopia. World Animal Review 41: 2-10.

 

Beyene Kebede 1992 Estimation of additive and non additive genetic effects for growth, milk yield and reproduction traits of crossbred (Bos taurus x Bos indicus) cattle in the wet and dry environment in Ethiopia. PhD. Thesis, Cornell University. Ithaca, New York. pp 235.