Twelve Boran and 12 exotic cows (Friesian and Ayrshire) with their newborn calves were used in an experiment to evaluate the effects of restricted suckling (RS) on growth rate of the calves and milk yield and reproductive rate of the cows compared to artificial rearing (AR). In each breed, the animals were randomly divided into two groups: RS and AR according to calving date. Cows and calves in both groups were kept under the same conditions. Calves under AR treatment were removed from their dams 3 days after calving and fed milk from buckets. Calves under RS treatment were allowed to suckle their dams twice daily for 30 minutes after milking.

The average daily weight gain of the RS calves was higher (P<0.05) than that of AR calves (0.45 vs.0.40 kg/day). Milk conversion per kilogram weight gain was 3.9 kg for RS calves and 4.5 kg for AR calves (P<0.05). Average total milk yield of RS cows was 26% higher (P<0.05) than that of AR cows (9.1 vs.7.2 kg/day). Saleable milk of RS cows was 37% higher (P<0.01) than that of AR cows. There was no difference in postpartum interval to first oestrus (97 days for RS cows and 93 days for AR cows) or milk progesterone concentrations between cows in the two systems of calf management. Twenty five percent of the cows in AR treatment had mastitis while no case was observed in the RS cows. There were differences between breeds: Boran calves had higher (P<0.001) daily weight gain (0.46 vs.0.38 kg/day) and better milk conversion per kg weight gain (3.1 vs.5.3 kg) than exotic calves. Boran cows had lower daily milk yields than exotic cows (4.3 vs.12.0 kg), but gained 28 kg live weight in 6 months postpartum while exotic cows lost 9 kg.

Restricted suckling is considered to be a more appropriate system than calf removal and artificial rearing for managing cows and calves and can be recommended especially for small-scale farmers.

Cattle in developing countries are mainly Bos indicus and milk production is frequently based on dual purpose, or even triple purpose breeds, where the cows are also used for draught (Preston and Murgueitio 1992).

There are different ways of rearing calves, such as suckling ad libitum, different regimes of restricted suckling (RS) and artificial rearing (AR). Artificial rearing systems have been applied largely in industrialized countries. Calves are removed from their dams and cows are milked by machine, and then calves are bucket fed with milk from their dams or with a milk replacer. This system is frequently inappropriate under the conditions of feeding and management in most developing countries (Preston 1984). In addition, no matter how efficiently a cow is machine-milked, there is always residual milk left in the udder which cannot be utilized. This can be as much as 15% to 25% of the total milk (Bathet al 1985; Ugarte 1977). The incidence of mastitis and calf mortality is often higher under the AR system than under RS (Ugarte and Preston 1975; Knowles and Edwards 1983).

The restricted suckling system allows the calves to suckle their dams for a limited period after milking and the calves also use residual milk in the udder of the cows, which is milk not utilized in the AR system. According to a number of investigators (Ugarte and Preston 1972; Alvarez et al 1980; Knowles and Edwards 1983), RS increases growth rate of the calf, milk yield and saleable milk of the cow, while reducing the incidence of mastitis and has little or no negative effect on reproduction compared to AR.

Restricted suckling also appears to stimulate the cow's milk production and may increase the total amount of milk produced by the cow. Restricted suckling is often a more viable option than AR, as it reduces labour requirements, capital outlay and calf mortality (Preston and Leng 1987).

In Tanzania, farmers in some regions use a RS system for calf rearing, but other farmers and some state farms, including University and Research Centre farms, practice the AR system.

This experiment was undertaken to evaluate the effects of RS compared with AR on the growth rate of calves, milk yield and reproductive rate of Boran cows and exotic breeds. The objective was also to determine which of the two calf rearing systems could best be applied in conditions of smallholder farming.

The experiment was carried out at Sokoine University of Agriculture in Morogoro, Tanzania. Morogoro is situated 500 m above sea level with an annual rainfall of approximately 900 mm, falling mainly from April to September, and an average annual temperature of 25 °C

The experiment was conducted for 10 months. It was started in February 1993 and lasted until the end of December 1993.

Twelve Boran and 12 exotic cows (6 Friesian and 6 Ayrshire) with newborn calves were used in the experiment. They calved from February to June, 1993. The exotic cows were in their second or higher lactation, while the Boran cows, although not primiparous, had never been milked.

The experiment was designed with two factors: the classification factor (breed) and the treatment factor (calf rearing system). The animals were distributed over the two treatments: (1) Restricted suckling; (2) Artificial rearing; according to calving date and with the following procedure. For each breed (Boran and exotic breed), the first calving cow and calf were allocated to treatment (1), the next calving cow and calf were allocated to treatment (2), etc.

Table 1: Ingredients and chemical composition (%) of the concentrate supplied to the cows and calves in the experiment
	Cows	Calf
Ingredient
Maize bran	78	60
Cottonseed cake	20	38
Minerals	2	2
Chemical composition (% of DM)
N x 6.25	16	22
Crude fibre	5.9	7.2
Ether extract	8.9	7.6

During the experimental period all cows were grazed together on natural pasture and housed in the evening and fed the same basal diet of cut green grass, hay or silage given ad libitum. Exotic cows were fed 6 kg/day of a concentrate mixture (maize bran, cotton seed cake, and minerals; see Table 1); Boran cows received 2 kg per day. The concentrate was given at the time of milking. The cows were milked twice a day (6 am and 4 pm) and milk yield was recorded daily. Oestrous of the cows was detected daily in the morning, afternoon and evening by a teaser bull and visual observation of oestrous signs. Cows showing oestrous were mated naturally by a bull at the point of standing oestrous. Pregnancy was determined by rectal palpation 60 days after the service.

Calves in the AR groups were kept with their dams until three days of age. On the fourth day, cows and calves were separated and cows were milked by hand. Calves were kept in pens which contained an area of shelter, feeding and water troughs and a simple design for individual bucket feeding of milk through the wooden fence surrounding each enclosure. To facilitate milking of Boran cows, their calves were used to stimulate milk letdown by letting them suckle each of the cow's teats for a few seconds. They remained close to their dams during milking, but were not allowed further suckling. Calves were weaned at 6 months of age. During the six months before weaning, each Boran calf was fed a total of 270 kg milk and each exotic calf was fed 360 kg milk in declining amounts. Milk was given twice daily (morning and evening). Also, each calf of both breeds was fed a total of 200 kg concentrate (Table 1) in increasing amounts. In addition, calves received green grass and water ad libitum throughout the experiment.

Calves in the RS group were also separated from their dams three days after calving, when milking of the cows began. To stimulate milk let down the calves were allowed to suckle each teat of their dams for a few seconds. Cows were then milked by hand leaving one quarter of the udder plus the residual milk in the other quarters to be consumed by the calves after milking. After each milking, calves suckled their dams for 30 minutes. This milking system continued for 45 days of the lactation. After 45 days, all four quarters of the udder were milked and the calves left only with the milk obtained during letdown stimulation and the residual milk after milking. After suckling, calves were separated from their dams and kept in pens similar to the pens of AR calves. They were given the same amount of concentrates as the calves in AR groups (200 kg during 6 months in increasing amounts) and given free access to green grass and water throughout the experiment. The calves were weaned at 6 months of age.

Milk yield was recorded daily. Body weight of cows and calves at calving was recorded and subsequently they were weighed every 2 weeks throughout the experiment. Belly girth of the calves was measured every two weeks from 10 weeks of age throughout the experiment. Cow and calf health was monitored by recording any diseases and subsequent treatments given to the animals during the experiment.

Analysis of milk composition (protein and fat content) was done every two weeks. Nitrogen content was analyzed by the Kjeldahl method and fat content was analyzed by the Gerber method, following the procedure of Richardson (1985). The amount of milk consumed by the calves in the restricted suckling groups (weighed before and after suckling) was determined every two weeks. Milk samples were taken at 2-week intervals while the calves were being suckled. The samples were analyzed for protein (N x 6.25) and fat content.

Milk samples were taken twice weekly from day 10 after calving for determination of progesterone concentration. The aim was to monitor ovarian activity until the cows were confirmed pregnant by rectal palpation. A 10 ml milk sample was taken from the same quarter and the last stripping into a tube which had a solution of Bromopol and Methylene Blue. The samples were centrifuged at 3000 G for 15 minutes. After that the skim milk was aspirated and transferred into small plastic containers and stored in a deep-freezer (-20 ?C) until the day of analysis. Progesterone in skim milk was analyzed by the Radioimmuno Assay method. The Coat-A-Count progesterone kits used were from DPC (Diagnostic Products Corporation) with standards provided by the IAEA (International Atomic Energy Agency). The assay procedure followed the manual of the IAEA (Progesterone RIA Pre-coated tube method. Assay Protocol, Version 2.0, January 1993), using a single-well gamma counter. Progesterone concentrations were calculated by the RIAPC program (Radioimmuno Assay Calculation 1988). STATISTICAL ANALYSIS

The data were analyzed using the General Linear Model in Minitab statistical software version 8.2 (Minitab Inc., 7081 Enterprise Drive, State College, PA 1680-7008, USA). The analysis of variance was done for all production parameters and the effects of breed, treatment and interaction between breed and treatment were analyzed.

Mean value for growth rate of the calves are in Table 2. For birth weight of the calves, there was no significant difference between the management systems but a significant difference between breeds, with exotic calves having higher (P<0.05) birth weights than Boran calves (27.7 vs. 25.3 kg). There were significant differences in body weight at 6 months of age, with RS calves having higher live weight (P<0.05) than AR calves (109 vs. 99.7 kg) and Boran calves having a higher mean live weight (P<0.01) than exotic calves (111 vs. 97.6 kg). Average daily weight gain from birth to 6 months was 0.45 kg for RS calves and 0.40 kg for AR calves (P<0.05), 0.46 kg for Boran calves and 0.38 kg for exotic calves (P<0.01). Mean milk conversion per kg weight gain was 3.9 kg for RS calves and 4.5 kg for AR calves (P<0.05), 3.1 kg for Boran calves and 5.3 kg for exotic calves (P<0.01) (Table 2).

Table 2: Mean values of weight gain and milk conversion (kg) of the calves
	Liveweight		Daily	Milk	Milk
	Birth	26 weeks	gain	intake	conversion
Treatments
Boran RS	24.7	116	0.49	1.4	2.8
Boran AR	26.0	105	0.44	1.5	3.4
Exotic RS	27.2	101	0.40	2.0	5.1
Exotic AR	28.2	94	0.36	2.0	5.5
Main factors
Management
RS	25.9	108	0.45	1.7	3.9
AR.	27.1	99	0.40	1.8	4.5
Prob.	0.30	0.01	0.03	0.001	0.016
Breed
Boran	25.3	110	0.46	1.4	3.1
Exotic	27.7	97	0.38	2.0	5.3
Prob.	0.045	0.001	0.001	0.73	0.001
SE of mean	±0.77	±2.6	±0.014	±0.10	±0.14

Table 3: Average daily weight gain (kg) of the calves from birth through 26 weeks divided into two week periods
	Boran			Exotic			Calf rear		Breed
	RS	AR	RS	AR	RS	AR	Boran	Exotic	SE
Week
2	0.28	0.23	0.21	0.20	0.25	0.21	0.25	0.21	0.02
4	0.37	0.31	0.24	0.24	0.31	0.28	0.34	0.24	0.02
6	0.42	0.36	0.28	0.30	0.35	0.33	0.39	0.29	0.02
8	0.34	0.36	0.24	0.30	0.29	0.33	0.35	0.27	0.02
10	0.40	0.36	0.32	0.31	0.36	0.33	0.38	0.32	0.02
12	0.49	0.40	0.34	0.31	0.41	0.35	0.44	0.33	0.03
14	0.50	0.48	0.38	0.31	0.44	0.39	0.49	0.34	0.03
16	0.54	0.47	0.38	0.38	0.46	0.43	0.51	0.38	0.02
18	0.57	0.47	0.46	0.43	0.51	0.45	0.52	0.45	0.02
20	0.58	0.51	0.49	0.44	0.54	0.48	0.55	0.46	0.03
22	0.59	0.52	0.55	0.46	0.57	0.49	0.55	0.51	0.03
24	0.67	0.59	0.64	0.49	0.66	0.54	0.63	0.57	0.03
26	0.77	0.62	0.74	0.60	0.76	0.61	0.69	0.67	0.03

Table 3 shows the average daily weight gain of the calves during their first six months. In both treatments, average daily gain of the calves was low during the period soon after birth, but it increased as the calves became older. The RS calves had decreased daily weight gain at the 8th week of age. There was little difference between treatments during the early phase of life but this became more pronounced in the later phase.

Table 4: Mean values of body score (live weight/belly girth) of the calves from 10 to 26 weeks of age
	Boran		Exotic		Calf rear		Breed
	RS	AR	RS	AR	RS	AR	BR.	Exot.	SE
Week
10	0.57	0.53	0.52	0.50	0.54	0.52	0.55	0.51	0.011
12	0.59	0.57	0.55	0.53	0.57	0.55	0.58	0.54	0.012
14	0.63	0.61	0.58	0.55	0.60	0.58	0.62	0.57	0.014
16	0.66	0.63	0.59	0.58	0.63	0.61	0.65	0.59	0.013
18	0.71	0.66	0.63	0.61	0.67	0.63	0.69	0.62	0.016
20	0.75	0.70	0.66	0.64	0.70	0.67	0.72	0.65	0.015
22	0.79	0.74	0.70	0.67	0.75	0.70	0.77	0.69	0.015
24	0.83	0.78	0.76	0.71	0.79	0.74	0.80	0.73	0.016
26	0.87	0.80	0.80	0.74	0.84	0.77	0.84	0.78	0.015

Restricted suckling not only affected growth, but also had an effect on the body conformation of the calves. A score was calculated (body weight/belly girth) and is presented for data from 10 to 26 weeks of age (Table 4). The data show that during every period RS calves had a higher score than AR calves. The average score was 0.68 for RS calves and 0.64 for AR calves. Similar to the difference between treatments for average daily weight gain, the difference between treatments for body score was lower in the early period and higher during the later period. Boran calves had a higher score than exotic calves, the average score being 0.69 for Boran and 0.63 for exotic calves.

There was a significant difference between treatments (P<0.05) with regard to milk yield. The cows on restricted suckling had an average daily milk yield which was 26% higher than that of AR cows (9.1 kg for RS and 7.2 kg for AR cows). The exotic cows had higher (P<0.01) daily milk yield than Boran cows (12.0 kg vs. 4.3 kg). Average milk consumed daily by the RS calves was 1.7 kg, this being 19% of total milk produced. Saleable milk was significantly different between treatments and breeds (P<0.01), RS cows having 37% more saleable milk than AR cows (7.4 kg vs. 5.4 kg) and exotic cows having more than three times higher saleable milk than Boran cows (9.9 kg vs. 2.9 kg) (Table 5).

Table 5: Mean values of daily milk yield and saleable milk (kg) of the cows
	Milking	Calf	Total	Saleable
Treatment
Boran RS	3.7	1.4	5.1	3.7
Boran AR	3.5	1.5	3.5	2.0
Exotic RS	11.1	2.0	13.1	11.1
Exotic AR	10.8	2.0	10.8	8.8
Main factors
Management
RS	7.4	1.7	9.1	7.4
AR	7.2	1.8	7.2	5.4
Prob.	0.77	0.012	0.01
Breed
Boran	3.6	4.3	2.9
Exotic	10.9	12.0	9.9
Prob.	0.001	0.001	0.001
SE	±0.48	±0.49	±0.48

Table 6: Change in live weight postpartum (kg) of the cows
	At	At 26
	calving	weeks	Difference
Treatment
Boran RS	309	337	28
Boran AR	331	359	28
Exotic RS	406	401	-5
Exotic AR	410	397	-13
Main factors
Management
RS	357	369	11
AR	371	378	7
Prob.	0.58	0.79
Breed
Boran	320	348	28
Exotic	408	399	-9
Prob.	0.02	0.032
SE	17	16

There was no difference between treatments in change of live weight postpartum. There was a difference between the breeds (Table 6), Boran cows increasing by 28 kg while exotic cows lost 9 kg.

There were no differences in milk protein between normal milk and that sucked by the calves (residual milk). However, a difference was observed between treatments in normal milk whereby AR cows had a higher (P<0.01) protein content than RS cows (3.33% vs. 3.14%). Fat content showed a slight difference between treatments (3.31% for RS cows and 3.49% for AR cows), but there was a significant difference between normal and suckled milk, milk sucked by calves having a higher fat content than normal milk (5.90 vs. 3.31%). There were also a slight difference between breeds for milk composition; the Boran cows had a higher protein and fat content in normal milk and protein content in suckled milk, while the exotic cows had higher fat content in suckled milk (Table 7).

Table 7: Mean values of protein and fat content (%) of normal milk and residual milk (milk sucked by calves)
	Protein		Fat
	Normal.	Resid.	Normal	Resid.
Treatment
Boran RS	3.17	3.19	3.36	5.78
Boran AR	3.36	3.61
Exotic RS	3.12	3.15	3.26	6.02
Exotic AR	3.30	3.36
Main factors
Management
RS	3.14	3.17	3.49	5.90
AR	3.33	3.31
Prob.	0.02	0.19
Breed
Boran	3.26	3.31
Exotic	3.21	3.49
Prob.	0.31	0.20
SE	±0.05	±0.12

The data in Table 8 show that in both breeds there was no clear effect of treatment on postpartum interval to first oestrous, but the RS cows showed a slightly longer interval than the AR cows (97 vs. 93 days). Boran cows had longer interval to first oestrous postpartum than exotic cows (103 vs. 88 days).

Table 8: Mean values for days postpartum to oestrous, days with progesterone in milk >3.2 nmol/litre, progesterone concentration (nmol/l) in skim milk and mastitis incidence (%) of the cows
	Postpartum.	Prog.	1st	Mastitis
	to oestrous	>3.2nmol	Prog.	incidence
	(days)	(days)	(days)
Treatment
Boran RS.	97	89	21.3
Boran AR.	109	97	20.7	33.3
Exotic RS.	97	71	21.2
Exotic AR.	78	70	22.3	16.7
Main factors
Management
RS	97	80	21.2
AR.	93	84	21.5	25
Prob.	0.69	0.69	0.92
Breed
Boran	103	93	21.0
Exotic	87	71	21.7
Prob.	0.11	0.03	0.78
SE	±6.54	±6.82	±0.98

For RS and AR cows, progesterone appeared (>0.2 nmol/litre) 78 and 82 days after calving, but the interval until progesterone concentrations exceeded 3.2 nmol/litre was 80 and 84 days, for RS and AR cows respectively. Boran cows had longer intervals than exotic cows (91 vs. 70 days for appearance of progesterone [>0.2nmol/litre] and 93 vs. 71 days for when progesterone concentration exceeded 3.2 nmol/litre). The maximum progesterone concentration during the first cycle not preceded by oestrous was 21.2 and 21.5 nmol/litre, while the duration of existence of luteal concentrations of progesterone (>3.2 nmol/l) were 11 and 10 days, respectively, and there was no difference between breeds. In the normal regular cycle preceded by oestrous behaviour, maximum progesterone concentration and duration of luteal phase progesterone concentrations were similar to those of the first cycle, and there was also no difference between breeds (Table 8).

There were three cases of clinical mastitis observed in the AR groups, one among the exotic cows and two among the Boran (25% of total AR cows). No incidence of clinical mastitis was observed in RS cows (Table 8).

Higher rates of weight gain and better milk conversion per unit weight gain of the RS calves in the present study could be related to the higher fat content in the milk consumed by those calves compared to the milk consumed by AR calves and also to the advantages for the digestive tract of young calves when they suckle directly and slowly from their dam. When calves are suckled naturally, milk is channelled by the reticular groove reflex directly to the abomasum, which is the true stomach of a newborn calf. Both the energy and protein are utilized with considerably greater efficiency than if the milk had first passed into the rumen. When drinking milk from a bucket, it is probable that a part of the milk enters the rumen. The Boran calves had higher daily weight gain and better milk conversion than the exotic calves, which is probably due to them having better adaptability to local conditions and a higher efficiency of milk utilization.

Alvarez et al (1980), working with crossbred European/Zebu cows, showed that RS calves had growth rates higher than calves reared with the AR system (446 g/day vs. 277 g/day). Little et al (1991) studied the influence of RS in crossbred dairy cows (Friesian x Boran) and showed that RS increased calf growth rate to weaning by 0.22 kg per day. Gonzalez et al (1984) found that in Friesian calves the daily weight gain was 440 and 598 g for AR and RS calves. Carias and Vaccaro (1984) studied the two systems of calf rearing in Holstein-Friesian and Brown Swiss cattle and observed that the restricted suckling system produced heavier (P<0.05) weight gains than the bucket system (0.677 vs. 0.555 kg/day).

Alvarez et al (1980) and Velazco et al (1983) demonstrated that growth rates of calves and conversion of milk into live weight gain were improved when calves were reared by RS rather than with milk from a bucket. Knowles and Edwards (1983) compared the effects of the RS and AR systems on dam and calf performance and showed that with the RS system, growth rate of the calf was 34% higher, while milk consumed for one kg of weight gain was 22.5% less during 70 days. These results are also similar to data reported by Ugarte (1978) and Fernandez et al (1977).

The data for growth rate of the calves during the present experiment show that, in general, calves on both treatments grew at acceptable rates. Growth rate of all calves increased as the calves grew older compared to earlier periods. Most calves had a lower average daily gain during the period from 8 to 10 weeks, and thereafter it increased. The reason for this is that the calves were dewormed after 2 months. Also, calves in RS groups had lower daily weight gain in the 8th week due to reduction in milk consumed, as from 45 days after birth they sucked only residual milk after milking. The difference between treatments was larger during the later periods.

Data on body score (live weight/belly girth) of the calves indicate that RS calves had better body condition than AR calves. The AR calves were thinner and had a bigger belly than RS calves at the same body weight. It is often so that at the same body weight, the fatter calves have a smaller belly and thinner calves have a bigger one. This could be related to the large amount of milk which the AR calves were fed, of which a part of the milk probably entered the rumen and was fermented, and possibly caused distension of the rumen. Ugarte et al (1974) indicated that with the RS system, calves had significantly higher carcass weight, yield and excess fat and less bone than those reared with nurse cows.

Milk yield in RS cows was higher than in AR cows, which confirms the high efficiency of utilization of the cows' milk potential when milking is combined with suckling (Ugarte and Preston 1973). After milking, RS calves were suckled for 30 minutes, and got residual milk which is not utilized in the AR system. Another reason is that when the calves are suckled by their mother, the brain of the dam is stimulated and oxytocin continues to be secreted and affects the milk ejection. Milk yield of the exotic cows was higher than in Boran cows, which confirms the high potential of milk production of European breeds specialized for milk production.

Paredes et al (1981), Alvarez et al (1980), Ugarte and Preston (1972, 1975) and Gaya et al (1977) studied the effects of RS on milk yield of cows and showed that compared with AR methods, RS prolonged the lactation and increased the milk yield. Velazco et al (1982) showed that Holstein cows in Venezuela, milked by machine and with restricted suckling of their own calves after milking, gave more milk and lost less body weight after calving than cows whose calves were removed after three days and reared artificially. Little et al (1991) showed that milk yield was 15% higher for RS cows compared with those whose calves were artificially reared. Knowles and Edwards (1983) also compared the RS and AR systems and reported that milk yield was higher (2615 litres vs. 1960 litres in 305 days of lactation, respectively), and saleable milk was 69% higher in the RS system. Teeluck et al (1981) studied Creole and Creole x Friesian cows and showed that compared to AR cows, RS cows had higher total daily milk yields (6.37 vs. 8.55 kg) and saleable milk (3.06 vs. 6.39 kg). Ugarte and Preston (1973) found that the average increase in milk secretion attributed to suckling was 31% in RS systems.

Milk consumed by restricted suckled calves in this study was 19% of total milk. Ugarte (1977) reported that there is always residual milk left in the udder which can not be utilized and it was 21% of the total milk in his study.

There were no significant differences between treatments in change of liveweight of the cows. Similar results, reported by Gaya et al (1977), showed that there were no significant differences between RS and AR in live weight changes of the cows. Ugarte and Preston (1975) reported that live weight of dams changed from birth to weaning of calves (at 70 days), with a decrease of 17 kg for RS cows and an increase of 2.8 kg for AR cows. In the present study, there were differences between breeds and the Boran gained weight while the exotic breeds lost body weight postpartum. This may be due to the differences in nutrient partitioning between Bos taurus and Bos indicus cattle. Also, the exotic cows may have lost body weight due to the poor adaptation in the tropics of European cattle breeds specialized for milk production. This poor adaptation creates management difficulties due to the need to protect them against climatic stress, as well as the high level of management and feeding required by European breeds.

Protein and fat content of the normal milk from RS cows was lower than that of AR cows, which could be due to the fact that RS cows had higher milk yield and therefore the protein and fat content was lower. A slight difference between breeds in protein and fat content could be due to differences of genotype. Fat content in the residual milk, which was sucked by calves, was much higher than that in normal milk. It is one of the important reasons explaining the higher growth rate of the calves reared in the RS system, although they consumed less milk compared to calves reared in the AR system. In the study by Ugarte (1977), similar results were obtained showing that fat content in normal milk with machine milking was 2.8% while the last stripping milk had 5.7% fat. These results are in agreement with data published by Johnson and Alford (1974) and Topel (1976).

In this experiment almost all cows had an ovulation without oestrous being detected. It could be explained by a physiological adjustment whereby the animal may defer acceptance of the bull until the time that she is physiologically and physically capable of conception and maintenance of pregnancy (Tegegne et al 1993). There were no differences between treatments in the interval postpartum to first oestrous. Little et al (1991) found that RS delayed return to oestrous (101 vs. 41 days), but required fewer services per conception (1.3 vs. 1.9). However, Alvarez et al (1980) studied the effects of RS and AR on time to first oestrous and calving interval of the dam and showed that RS had no or little effect on reproductive rate. Veitia and Simon (1972) studied F1 Holstein x Brahman cows and reported that there were no differences in time from calving to first oestrous for RS and AR cows (84 and 81 days respectively). Ugarte and Preston (1975) worked with Holstein cows and observed that the postpartum interval from calving to first oestrous was 89.6 and 88.5 days for RS and AR cows. A study by Leon and Vaccaro (1984) even showed that RS twice daily decreased the calving interval and services per conception, while milk yield of the cows and growth rate of calves increased compared with AR and RS once daily systems. Knowles and Edwards (1983) reported that there was no effect on the reproductive efficiency of the suckled and non-suckled cows. Ferreira and Torres (1991) studied the effect of restricted suckling on postpartum ovarian activity in Holstein x Zebu heifers and found that there were no significant differences between RS cows and AR cows in the interval from parturition to first oestrous (94.4±5.8 and 88.5±8.2 days respectively). Montoni and Riggs (1978) showed that for Brahman cows suckled twice daily compared with cows suckled ad libitum, the interval from calving to first oestrous was reduced.

The duration of the luteal phase of progesterone concentration in the milk in this experiment appeared to be short, because samples were taken twice a week, so the interval from sample to sample was 3.5 days. In reality if the luteal phase extended over days when no samples were taken the length could be from 13 to 15 days. The exotic cows had a shorter interval from calving to oestrous and progesterone appearance, which may be due to the effect of selection for this trait in Bos taurus animals. The progesterone concentration and the duration of luteal progesterone concentration during the first cycle which was not preceded by oestrous were normal, and there were no differences between breeds or treatments.

Caruthers et al (1980) found that the serum concentration of progesterone did not differ between suckled and non-suckled cows. Suzuki and Sato (1981) studied the effect of restricted suckling on postpartum reproductive performance of Japanese Black cows compared to suckling ad libitum and proved that there were no significant differences between groups for cow serum progesterone concentrations, interval from calving to first ovulation, service period or the number of cows showing oestrous at the first ovulation. In other studies (1983 and 1984) they indicated that RS tended to shorten postpartum intervals to first ovulation, first oestrous and conception compared to normal suckling. Bastidas et al (1984) studied the effect of RS on ovarian activity and found that the average interval from parturition to first corpus luteum was 59.0±2.0 days and to first oestrous was 68.0±5.0 days. Velazco et al (1982) found that cows which were suckled under a restricted regime had lower peak concentrations of blood progesterone, when compared to those cows which were not suckled. They suggested that low fertility in restricted-suckled animals could possibly be caused by a high incidence of "silent" or quiescent oestrous, due to low progesterone. Mukasa-Mugerwa et al (1991) investigated the ovarian activity as determined by plasma progesterone concentrations of >3.9 nmol/l and indicated that "silent ovulation" was initiated in 44% of cows by 40.6 days postpartum, while the mean interval from calving to first behavioral oestrous was 81.1 days. Suzuki and Sato (1983) also determined that the interval from calving to first ovulation ranged from 20.4 to 28.6 days and from calving to first overt oestrous was from 23.6 to 44.4 days in three groups of experimental cows. Tegegne et al (1993) studied short luteal phases and ovulations without oestrous in primiparous Borana cows and found that almost all cows had at least one regular oestrous cycle prior to the first observed oestrous.

Since, in this experiment the first oestrous cycle postpartum of both normal length and adequate progesterone concentrations was independent of breed, it is possible, and can be recommended, to breed these cows at first oestrous postpartum.

There was no mastitis problem in RS cows, which can possibly be attributed to the greater stimulus for evacuating the udder when these cows suckle their calves. Alvarez et al (1980) reported that mastitis incidence was reduced when cows were milked with the calves reared by RS. These results are in agreement with findings of Ugarte and Preston (1975).

The restricted suckling system improved growth rate of the calves, increased milk yield and saleable milk of the cows, and reduced mastitis incidence, but had no effect on the reproductive rate compared to the artificial rearing system.

Restricted suckling is therefore a more appropriate system for managing both cows and calves, than is calf removal and artificial rearing, and can be recommended especially for smallholder farmers.

The authors gratefully acknowledge the financial support from the Swedish International Development Authority (SIDA) and the cooperation of Sokoine University of Agriculture, Tanzania (SUA) for this study. The authors would like to thank Professor F Lekule for his advice and organizing of the experiment, and Mr S Sarwatt for his collaboration. Thanks are also due to Mr J Maimbi, Mr J Chugga, Dr Nkya, Mr M Kalimba, Mrs M Mbwana and Mr Y Watuta for their assistance.

(Received 1 July 1994)