4. REPRODUCTION

4.1 Reproduction of Barbary ewes

The seasonal character of sexual activity in sheep has long been known. Among the factors controlling seasonal reproduction, photoperiod is without doubt the most important component (Yeates, 1949; Hafez, 1952; Menaker, 1971; Follett, 1978; Legan and Karsh, 1980; Goodman and Karsh, 1981; Thimonier, 1981).

Generally, cyclic oestrous activity appears in ewes when the daily clear photoperiod decreases and seasonal anoestrus occurs when this photoperiod begins to increase (Ortavant et al. 1964; Fraser and Laing, 1969; Thimonier and Mauleon, 1969; Ducker and Bowman, 1970; Newton and Betts, 1972; Dyrmundsson, 1978; Lax et al. 1979). The traditional mating season of Barbary females in Tunisia is in spring (Tchamitchian and Sarson, 1970; Sarson, 1972; Khaldi, 1980). Reasons to explain this contradiction could be:

It is therefore essential to know the seasonal variations in the oestrous and ovarian activity of Barbary females in their natural environment to specify adequate reproduction periods.

On the other hand, in a favourable environment, where nutrition does not represent a limiting factor, the acceleration of lambing rate can be an efficient means to increase the number of lambs produced by ewes every year, but these intensive techniques depend on the aptitude of the female to become pregnant rapidly after parturition.

Many studies have shown that the resumption of ovarian and oestrous activities during the post-partum period can be affected by the lambing season, lactation, nutrition and breed (Mauleon and Dauzier, 1965; Hunter, 1968; Restall, 1971; Joubert, 1972; Shevah et al. 1974; Restall and Starr, 1977). After the lambing season, Hunter (1968) concluded that the nutritional level of ewes is an important factor which could affect their post-partum fertility. Rhind et al. (1980) came to a similar conclusion. Likewise, our own observations (Khaldi, unpublished data) showed a great variation in fertility rate of Barbary ewes managed in a 3 lambing/2 year system, especially when females were mated in December-January (35 to 85 percent). In this case, and for different reasons, feed availability was sometimes limited in autumn and winter, which led to more or less severe under-nutrition of females in late pregnancy and in lactation.

A series of experiments was thus performed to study the reproductive aspects of Barbary females.

4.1.1 Seasonal variations of ovarian and oestrous activity

A trial was performed at the Experimental Station of Bou-Rebiaa to specify the age at puberty and the seasonal variations of ovarian and oestrous activity of Barbary females. It started with 26 5-month old ewe-lambs and 25 4 to 5-year old ewes. The mean liveweight varied from 25 to 52 g in the ewe-lambs and from 41 to 62 g in the ewes during the 16 months of the experiment. Oestrus was checked twice daily and ovarian activity was controlled by regular coelioscopies every 17 days.

4.1.1.1 Breeding season and seasonal anoestrus

The onset of the breeding season of the ewe-lambs born in autumn occurs during the first half of September of the following year when they reach puberty at 10 months old weighing on average 34.7 g. This latter value represents about 65 percent of the adult ewe weight. This weight at puberty is a common characteristic in most animal species. The duration of the breeding season of ewe-lambs is short (104 days) since their last oestrus occurs in early December (Figure 15).

The breeding season of adult ewes is significantly (P < 0.01) longer. Its mean duration is 242 days and it extends from mid-July to late February (Figure 16). Evidence exists which indicates that the intensity of seasonal anoestrus is also related to female age. In effect, 25 to 40 percent of adult ewes ovulate and exhibit oestrus regularly in spring when, during the same period, the termination of the oestrous activity in young females is almost complete.

4.1.1.2 Ovarian activity

As with oestrous activity, cyclic ovarian activity is not a continuous phenomenon throughout the year, but is characterized by its seasonality and follows closely the changes of the former.

Irrespective of age, the results of this experiment show the existence of an oestrus-ovulation dissociation during some periods of the year. Thus, during a period of 12 months, 32.1 percent of ovulations are not accompanied by oestrous behaviour in ewe-lambs. The percentage of silent ovulations during the same period is only 22.6 percent in adult ewes. This phenomenon of silent ovulation is not peculiar to the Barbary breed, since it has been reported in many other breeds by Thimonier and Mauleon (1969), Van Niekerk (1972), Land et al. (1973), Hulet et al. (1974), Dyrmundsson (1981) and Thimonier (1981). Silent ovulation occurs throughout the four seasons but especially before the onset of the sexual season, after its termination and during April-May.

In contrast, the occurrence of oestrus without ovulation is very rare since the frequency of such a phenomenon represents only 2.6 percent of the total number of oestrouses recorded during all the experimental period.

The mean ovulation rate of the ewe-lambs is low (1.08) and remains relatively constant throughout the year (Figure 17). In contrast, the ovulation rate of adult ewes undergoes important seasonal changes with the highest ovulation rate being observed in September-October (1.60) and the lowest in March-April.

When a period of 12 months is taken into consideration (1 July - 30 June), the mean ovulation rate is significantly (P < 0.05) higher in ewes (1.32) than in ewe-lambs (1.08). On the other hand, this ovulation rate decreases in the case of silent ovulation, in both adult and young females, from 1.38 and 1.12 when they display oestrus, to 1.16 and 1.05.

4.1.1.3 Duration of oestrus and oestrous cycle

The mean duration of oestrus is estimated at 26.2 hours in ewe- lambs and 28.3 hours in adult ewes. The difference between the two categories of females is significant (P < 0.05). In both cases, the duration of oestrus is lower in spring than in summer and autumn (Figure 18). On average, the oestrus duration is longer more especially as the ovulation rate is higher (Table 10).

Table 10: Relationship between ovulation rate and oestrus duration (hours)

Source: Khaldi (1984).

The mean interoestrous interval is estimated at 17.7 days in both young and adult females and seasonal changes in oestrus cycle do not occur.

4.1.2 Post-partum anoestrus

4.1.2.1 Effects of suckling duration and lambing season

To study the effects of suckling duration and lambing season on the resumption of post-partum ovarian and oestrous activities, an experiment was performed on 131 adult females, 3 to 5 years old, at the Experimental Station of Bou-Rebiaa (Khaldi, 1984). Fifty-three ewes had lambed between 14 and 30 October, 35 between 12 and 26 February and 43 between 2 and 28 June. For each of these lambing periods, the mean prolificacy rate was 120, 140 and 126 percent respectively. The mean liveweight of ewes 48 hours after parturition was about 50 g.

In each lambing season, weaning of lambs was accomplished in half of the ewes 45 or 90 days post-partum except for October lambing ewes where a third group dried off only 2 days after parturition.

Animals were grazed on natural pastures in autumn, winter and spring and on cereal stubble in summer. In addition, they received hay ad libitum and 300 g of concentrate/ewe/day. Ovarian activity was controlled by coelioscopy every 17 days from the 5th day post-partum and oestrus was checked twice daily for 8 months.

4.1.2.1.1 Effects of suckling duration

Irrespective of the lambing season, the suckling duration (45 or 90 days) had no significant effect on the post-partum resumption of ovarian activity of ewes. On the other hand, drying off of autumn lambing ewes 48 hours after parturition did not lead to a significant decrease in interval between parturition and the occurrence of the first post-partum ovulation (Table 11). In this case, the intervals were 15.1, 16.5 and 17.6 days when the females dried off 2, 45 or 90 days after lambing.

Table 11: Effects of lambing season and suckling duration on parturition post-partum first ovulation interval

Source: Khaldi (1984).

In a given lambing season, suckling duration (45 or 90 days) was of no consequence on those ewes displaying at least one oestrus during the 8 months following parturition and the mean interval between parturition and the occurrence of the first oestrus in these ewes during the same period (Table 12). In contrast, the precocious drying off of ewes (48 hours after lambing) in October reduced their post-partum period by about 25 days.

Table 12: Percentage of ewes displaying at least one oestrus within the first 240 days post-partum and parturition-oestrus interval (days)

Source: Khaldi (1984).

4.1.2.1.2 Effects of lambing season

As the resumption of cyclic oestrous activity during the post- partum period was not affected by suckling duration (45 or 90 days), the effect of lambing season on the interval between parturition and first oestrus was calculated for all the ewes of each group (Table 12). The average values of the post-partum anoestrus duration (50 percent of females displaying oestrus) were 55, 68 and 108 days in ewes lambing in October, June and February respectively. The percentage of females presenting at least one oestrus during the experiment (8 months) was 90.3 in the first group, 85.7 in the second group and 80.0 in the third group (P < 0.05).

4.1.2.2 Effects of nutritional level on pregnancy and lactation

The influence of the pre-partum and post-partum nutritional levels of Barbary ewes was studied by Khaldi (1984) at the Experimental Station of Bou-Rebiaa. The experiment concerned 164 adult females mated between 12 and 19 May after oestrous synchronization (30 mg of FGA and 400 IU of PMSG). At the start of the trial, the mean age and liveweight of ewes were 4.6 ± 1.3 years and 55.5 ± 5.3 g. All the females were housed and they were divided into 4 groups of 41 ewes according to their nutritional level (Table 13) during the last 12 weeks of pregnancy and the first 18 weeks of lactation:

Table 13: Quantities of feed distributed to ewes (g/ewe/day)

Source: Khaldi (1984).

All the ewes were weighed every week and 24 hours after parturition. Maximum tail perimeter was also recorded at the beginning of the experiment, at parturition, and 4 and 18 weeks post- partum. The lambs were weighed at birth and then every week.

The post-partum ovarian activity was controlled by coelioscopy every week and the dosage of the plasmatic progesterone level in blood samples 3 times/week until the occurrence of the second post-partum oestrus for each ewe. Oestrus was checked twice daily be entire rams.

4.1.2.2.1 Feed intake

The mean quantity of hay dry matter (DM) ingested during the last 12 weeks of pregnancy was 1.30 g/ewe/day in the HH and HL groups. Its weekly variation was negligible. The ewes of these two groups consumed all the concentrate distributed (400 g/ewe/day). Groups LH and LL consumed only 0.87 g of hay DM ewe/day during the same period. Total energy intake of the underfed females before parturition represented only 44.4 percent of the well-nourished ewes.

During lactation, and over and above the 800 g of concentrate, the ewes of groups HH and LH consumed on average 1.56 and 1.47 g of hay DM ewe/day. These intake levels did not vary considerably during the 18 weeks of suckling. The underfed females of groups HL and LL consumed only 200 g of soya cake and 0.87 g of DM of hay/animal/day Their energy intake represented only 42.5 percent of consumption of well-nourished ewes.

4.1.2.2.2 Liveweight changes

The pre-partum nutritional level of ewes had a highly significant (P < 0.01) effect on the evolution of their liveweight during the last 12 weeks of pregnancy. In fact, the liveweight of the well-fed ewes (HH and HL) increased dramatically (9.4 g) until parturition while the underfed females (LH and LL) kept a constant liveweight during the same period in spite of the growth of their pregnant uterus. Thus, the latter lost at least 6.50 kg of their own liveweight (liveweight 24 hours after lambing - liveweight 12 weeks before lambing) (Figure 19).

After parturition, the principal factor influencing liveweight changes in suckling ewes was certainly their post-partum feed intake level. The well-nourished females did not lose much body mass (HH) or kept a constant liveweight (LH) during the first 18 weeks of suckling. On the other hand, the weight loss of the under-nourished females during the same period was significant since it represented between 21 and 29 percent of their liveweight 24 hours post-partum in groups LL and HL, respectively.

4.1.2.2.3 Caudal perimeter changes

During the last 12 weeks of pregnancy, the overnourished ewes (HH and HL) maintained a constant caudal perimeter but this perimeter decreased from 3.4 to 6.0 cm during the same period in the underfed animals (Table 14). Further, the effect of post-partum nutritional level on the caudal perimeter during lactation was highly significant (P < 0.001). This perimeter showed a reduction of more than 24 cm in the underfed females of groups LL and HL which lost about 50 and 43 percent respectively of the perimeter of their fat-tail during the suckling period.

Table 14: Effect of nutrition on the evolution of the caudal perimeter (cm) of Barbary ewes

Source: Khaldi (1984).

Irrespective of the physiologial stage of the ewes, there was a highly significant (P < 0.001) correlation between their caudal perimeter (Y, cm) and their liveweight (X, g). Nevertheless, the correlation coefficient (r) between these two criteria was higher at the end of lactation than at parturition or during the prior 12 weeks. Regressions obtained at these three stages were the following:

The correlation between caudal perimeter variations (Y, cm) and proper liveweight changes (X, g) during the last 12 weeks of pregnancy or the first 18 weeks of lactation were also very significant (P < 0.001). Reressions of the two criteria are:

4,1.2.2.4 Ovarian activity

Irrespective of litter size, pre-partum nutritional level had no significant effect on the date of first post-partum ovulation.

The effect of post-partum nutritional level on parturition-first ovulation interval was only significant (P < 0.05) in ewes suckling single lambs. In this case, the first post-partum ovulation occurred on average 9 days sooner in groups HH and LH than in groups LL and HL (Table 15). The resumption of ovarian activity of twins suckling ewes took place about 30 days after parturition in the 4 groups. Interval parturition first ovulation was significantly (P < 0.05) affected by the number of suckled lambs only in group HH where it was shorter in single mothers than in twin mothers.

Table 15: Effect of nutrition and number of suckled lambs on parturition first post-partum ovulation interval (days)

Source: Khaldi (1984).

The first post-partum ovarian cycle was not always of a normal duration (Figure 20). For all ewes, this duration varied from 5 to 79 days. Only 49.6 percent of the first ovarian cycle had a normal duration ranging between 15 and 20 days (17.2 ± 1.21 days). The first post-partum ovarian cycle was short (8.1 ± 2.5 days) or long (37.2 ± 17.1 days) in 30.6 and 19.8 percent of ewes respectively. The incidence of abnormal cycles occurred more frequently when the resumption of ovarian activity was pecocious. In fact, the duration of post-partum interval to ovulation was significantly (P < 0,05) longer in ewes with a normal cycle (30.0 ± 18.8 days) than in those with an abnormal cycle (23.0 ± 16.0 days).

The most important factor influencing the duration of the first ovarian cycle was the post-partum nutritional level of ewes. The frequency of short cycles was significantly (P < 0.001) higher in the well-nourished females of groups HH and LH (about 22 percent) than in the underfed ewes of groups LL and HL (about 13 percent).

Frequency of abnormal corpus lutea decreased significantly (P < 0.001) at the second post-partum ovulation. Thus, most females (83.8 percent) had a second ovarian cycle of a normal duration (17.9 ± 1.2 days). The

premature regression of the second corpus luteum was observed in 13.5 percent of ewes and its persistence in only 2.7 percent.

4.1.2.2.5 Oestrous activity

Resumption of post-partum oestrous activity was less precocious than ovarian activity since the first ovulation was silent in 99 percent of ewes. On the whole, the number of silent ovulatiHons ranged from 0 to 5 with a mean of 1.5 ± 0.8. This number was significantly (P < 0.001) affected by post-partum nutritional level. It was higher in the well-nourished ewes after lambing (HH and LH: 1.3 to 3.3) than in the underfed females during the post-partum period (LL and HL: 1.1 to 1.2). The prepartum nutritional level and the number of suckled lambs had no significant effect on the frequency of silent ovulation.

The occurrence of first oestrus depended upon the life span of the previous corpus luteum since 91 percent of oestrus was displayed after a normal or a long luteal phase and in 9 percent only of ewes with a preceding short luteal phase.

The percentage of ewes displaying at least one oestrus during the first 18 weeks of lactation depended essentially upon their pre- partum nutritional level. This percentage was significantly (P<0.05) higher in ewes receiving a high feed level in pregnancy (HH and HL: 89 percent) than in those suffering from under-nutrition during the same period (LL and LH: 66 percent). In the same way, this percentage was significantly (P<0.05) affected by the number of suckled lambs. It was higher in ewes suckling single lambs (67 to 100 percent) than in those suckling twins (50 to 84 percent). The post-partum nutritional level had no significant effect on the occurrence of oestrus in ewes.

Post-partum interval to first oestrus was not significantly affected by the feed intake level before or after parturition. In return, this interval was influenced (P<0.01) by the number of suckled lambs (Table 16), being about 10 days shorter in single suckling ewes than in those suckling twins.

Table 16: Effect of nutritional and number of suckled lambs on interval parturition - first oestrus (days)

Source: Khaldi (1984).

The duration of the first oestrous cycle was not affected by the nutritional level before and after lambing or by the number of suckled lambs. This duration was normal (18.0 + 1.1 days) in 94.3 percent of females displaying at least two successive oestrus. The other cycles were short (6 to 7 days) or long (21 to 22 days). The second oestrus duration was always normal (18.2 + 0.8 days).

4.1.3 Response to ram exposure

In some animal species, the reproductive function of females can be modified by exposing them to males (goat: Shelton, 1960; sow: du Mesnil, du Buisson and Signoret, 1962; mouse: Whitten, 1958).

In most sheep, the sudden introduction of rams to previously isolated anoestrous females induces a synchronized appearance of oestrus with two peaks of activity around the 18th and 23rd days after being run with males (Schinckel, 1954; Fairnie, 1976). The use of the endoscopy technique (Thimonier and Mauleon, 1969) demonstrated that contact with rams causes the ewe to show silent ovulation within the first 4 days of teasing; the first induced ovarian cycle can be of a normal (17 days) or a short (6 days) duration (Oldham et_ al_., 1979; Knight et al_., 1981).

The application of teasing techniques can be useful in North Africa to obtain group lambing in the rainy season (autumn) when feed availability is important. Therefore, the response of Barbary females to rams according to age and nutritional level was studied (Khaldi, 1984).

4.1.3.1 Effect of age

An experiment was performed at the Experimental Station of Ousseltia on 160 adult ewes (5.1 ± 1.6 years) and 40 yearlings (1.5 ± 0.1 years). Adult females had lambed in autumn and had been dry for more than 75 days. The flock was kept on natural pastures and did not receive any supplementation. It had been completely isolated from rams for about 9 months. Entire rams were introduced in the flock on 5 May with a ratio males/females of 1/10. The occurrence of oestrus was checked twice daily until 10 July and females were mated only at their second oestrus. Ovaries of all females were examined by laparoscopy the day the males were introduced (Day 0) and 9 days later for those without any corpus luteum on Day 0.

4.1.3.1.1 Ovarian activity

The age of females had a high significant (P<0.01) influence on spontaneous ovarian activity before the introduction of rams into the flock. Effectively, about half the adult ewes (50.6 percent) were cycling spontaneously whereas 22.5 percent of yearlings only had corpus lutea the day the males were introduced.

The stimulation of ovarian activity of the non-cycling females by the introduction of rams was very intense. Ovulation was induced in practically all anoestrous ewes (97.5 percent) and in 74.2 percent of yearlings. The difference between the two categories of females was significant (P<0.01). The ram-induced ovulations seemed to occur during the first 4 days of contact with males (Figure 21).

Whatever the age of the females, the duration of the first ram-induced ovarian cycle was not always normal. The percentage of females with a short first ovarian cycle (about 6 days) was 23.4 percent in adult ewes and 34.8 percent in yearlings, but the difference between the two percentages was not significant. Ovarian activity was not controlled beyond the nineth day after introduction of rams; nevertheless, the dates of oestrous occurrence seemed to indicate that the corpus lutea formed after the short cycle persisted normally (Figure 21).

Ovulation rates were similar (1.22) in spontaneously cycling adult ewes and yearlings before teasing. In contrast, the ram-induced ovulation rate was significantly (P<0.05) higher in ewes (1.42) than in yearlings (1.09). The induced ovulation rate of adult ewes was higher (1.47) when their first ovarian cycle was of a normal duration than if this latter were short (1.22).

4.1.3.1.2 Oestrous activity

Eighty-seven percent of stimulated adult ewes showed oestrus before Day (D) 26. This percentage was 73.9 only in ovulating young females by the male effect, but the difference between the two percentages is not significant.

Irrespective of age of females, the ram-induced ovulation or that occurring after a short ovarian cycle was generally silent (without oestrus).

In adult ewes, 71.6 percent of oestrus was observed between D 13 and D 19, with a maximum frequency at D 17. This oestrus was related to the ovulation occurring after a first normal ovarian cycle. About 21 percent of oestrus occurred between D 21 and D 25. It coincided with the ovulation occurring after a first short ovarian cycle followed by a normal cycle. In this case, the oestrous peak activity was observed at D 22. The mean interval between the introduction of males and the occurrence of oestrus was 16.5 + 1.2 and 23.1 + 1.4 days in the two categories of ewes respectively.

Oestrus was observed in 52.9 percent of yearlings between D 16 and D 21 and in 47.1 percent between D 22 and D 24. Thus, the mean interval between the introduction of rams and the occurrence of oestrus was 18.0 + 1.6 in the former and 22.9 ± 0.7 days in the latter.

Sexual receptivity was observed at least twice in 82 percent of spontaneously cycling adult ewes before the introduction of rams into the flock and in 43 percent only in those ovulating by the ram effect. The other ewes had fallen again into anoestrus.

All ewes cycling before stimulation by the presence of rams had a normal oestrous cycle (17 days), but this duration was normal in less than 80 percent of the ram-stimulated ewes.

4.1.3.2 Effect of nutrition

This experiment studied the response of anoestrous Barbary ewes to ram effect according to their liveweight and their feed level between weaning of lambs and introduction of males to the flock (Khaldi, 1984).

The trial took place in the Experimental Station of Bou-Rebiaa. It involved 122 dry ewes of 3 to 6-years old (mean age: 4.1 ± 1.1 years). Nine weeks before rams were introduced (9 May), lambs ere weaned and ewes were divided into three lots according to their liveweight:

A different feeding ration was therefore distributed to the ewes to keep their liveweight constant, to decrease or to increase it before the introduction of rams. Five groups were formed:

The amount of distributed and refused feed was controlled daily for each group and the amount of forage was adapted weekly according to the mean liveweight change of each group.

Ovarian activity during the 21 days preceding the introduction of males was controlled by estimating the progesterone level in ewe plasma with a frequency of 3 samples/week. After the introduction of rams (Day 0), ovarian activity was controlled by laparoscopy at day 4, day 9 and 4 to 7 days after the occurrence of oestrus. This latter was checked twice daily with 4 entire rams per group and females were mated 12 and 24 hours after the onset of oestrus.

4.1.3.2.1 Feed intake and liveweight changes

Ewes of groups LH and HH consumed 400 and 200 g of concentrate/head/ day respectively. Quantities of hay infested were 1.2, 1.4, 1.0 and 1.1 g/head/day in groups LL, LH, HH and MM respectively. Mean intake of straw was about 0.8 g/head/day in group HL.

Liveweights of ewes remained practically constant in groups HH, LL and MM until the introduction of rams (Figure 22). Females of group HL, suffering from under-nutrition, lost 7.3 g ( - 116 g/day) between the onset of the experiment and the introduction of rams. In contrast, ewes of group LH gained livesweight (+ 105 g/day) during the same period. Their mean liveweight gain was 6.6 g.

Despite their very different body conditions at the onset of the experiment, the females of groups MM, LH and HL had a very similar liveweight (about 45 g) when males were introduced.

4.1.3.2.2 Ovarian activity

Initial liveweight at the onset of the experiment affected significantly (P<0.01) the proportions of ewes with spontaneously active ovaries before the introduction of males (LL and LH: 6.1 percent; HH and

HL: 40.8 percent). The percentage of cyclic females in group MM (12.5 percent) was not significantly different from those of the other groups. Liveweight changes between weaning and the introduction of rams (9 weeks) had no significant effect on the cyclic ovarian activity of ewes before the onset of teasing.

Ovulation was induced in most non-cyclic ewes of groups HH (88.5 percent), LH (91.3 percent) and MM (90.5 percent) within the first 4 days of contact with males. Ho' ver severe, more or less prolonged under-nutrition decreased the response of females to ram effect (Table 17). Thus, ovulation was induced only in 76.9 percent of females of group HL and 65.2 percent of females of group LL, but only the latter percentage was significantly (P<0.05) lower than that of groups LH and MM.

Table 17: Percentage of anoestrous ewes responding to ram effect

The ram-induced ovluation was followed in the quasi-totality of ewes by a second ovulation within the first 24 days of mating. Only 2 females of group HL did not reovulate during the same period, but the life span of their corpus lutea was normal. The first induced ovarian cycles did not always have a normal duration. Some females developed corpus lutea which regressed prematurely and thus had short Induced ovarian cycles (5.3 +0.7 days) but their subsequent ovarian cycle was of a normal duration (17.0 + 2.0 days) and similar to that of ewes in which the first induced ovarian cycle was normal (16.5 + 1.5 days).

The liveweight of ewes 9 weeks before teasing had a considerable influence on their induced ovulation quality. In effect, the short ovarian cycle frequency was significantly (P<0.01) higher in groups LL (53.3 percent) and LH (76.2 percent) than in groups HH (21.4 percent) and HL (20.0 percent). The percentage of ewes showing a short induced ovarian cycle in group MM was intermediate (31.6 percent).

In contrast, liveweight changes of ewes had no significant effect on the duration of their first ovarian cycle. On the other hand, whatever the liveweight changes, there was a threshold liveweight of 42.7g 9 weeks before teasing (calculated by discriminant analysis) by which the

proportion of females with a short induced ovarian cycle can be estimated. Under this threshold liveweight, 67 percent of females would have a short cycle. Above this threshold, the first induced ovarian cycle would be of a normal duration in 73 percent of ewes.

Liveweight of ewes 9 weeks before teasing did not seem to affect their induced ovulation rates (Table 18). In fact, they were very similar in groups LL (1.27) and HH (1.29). The ovulation rate of group MM was slightly higher (1.42). In contrast, the liveweight changes of females between drying off and the introduction of rams (9 weeks) had a dramatic effect on their induced ovulation rates. This effect appeared through the difference observed between groups LH (1.43) and HL (1.10). The influence of liveweight changes on ovulation rate of ewes at first oestrus was less clear. Nevertheless, the same tendencies concerning induced ovulation were found. This ovulation rate at first oestrus varied from 1.00 in group HL to 1.24 in group LH.

Table 18: Ovulation rate of stimulated ewes

Source: Khaldi (1984).

4.1.3.2.3 Oestrous activity

The occurrence of oestrus. in ewes ovulating by ram effect was not significantly affected by their liveweight and its changes. The percentage of females which displayed oestrus within the first 27 days of mating varied in fact from 80 to 100 percent according to groups.

4.1.3.2.4 Fertility and prolificacy

Only apparent fertility (number of lambing females per 100 females present in the flock at mating), real fertility (number of lambing females per 100 mated females) and prolificacy (number of born lambs per 100 lambing females) resulting from the first 27 days of mating were considered (Table 19). The liveweight of females had a marked influence on apparent fertility. This latter was significantly (P<0.01) lower in group LL (26.1 percent) than in groups HH (62.5 percent) and MM (66.7 percent). It was also affected by the liveweight changes before mating. In effect, the liveweight gain of group LH resulted in a significant (P<0.01) increase of

apparent fertility of ewes (78.3 percent) compared with ewes of group LL (26.1 percent). Likewise, the liveweight loss of ewes of group HL decreased their apparent fertility rate (38.5 percent) compared to the females of group HH. However, the difference between these two latter groups was not significant.

Table 19: Fertility and prolificacy of ram-stimulated ewes

Source: Khaldi (1984). !

The real fertility of ewes was equally affected by their liveweights. It was significantly (P<0.05) higher in group HH (83.3 percent) than in group LL (42.9 percent). That of group MM was intermediate (73.7 percent). In the same way, liveweight changes had an important effect on real fertility which was twice as high in group LH than in group LL (P<0.01). Moreover, liveweight loss of ewes in group HL decreased their real fertility rate by 25 percent compared with group HH (P<0.05).

Severe, prolonged (LL) or more recent (HL) under-nutrition of females affected their prolificacy. All ewes of groups LL and HL had single lambs. In the other 3 groups, the prolificacy varied from 110 to 116.7 percent. No significant difference was evident.

4.2 Reproduction in Barbary rams

The seasonality of semen production in Barbary rams was studied by Mehouachi and Khaldi (1987). Sperm was collected using an artificial vagina from 6 rams during a 13 month period with a rate of 2 x 2 collections per week. The results of this study showed that the Barbary rams ejaculated on average 11 x 10⁹ spermatozoa per week. Semen production was higher in summer and autumn than in spring.

The epididymal reserves seemed to be low as a result of decrease in sperm production between two ejaculates.

The volume of semen ejaculated by rams was significantly (P<0.001) affected by the season. Higher production was observed during summer (1.05 ml) and lower during winter (0.65 ml). The quality of ejaculated sperm was

also greatly influenced by season. However, variation of different parameters controlling this quality was not the same during the year:

The Barbary rams were sexually active throughout the year, but their libido decreased strongly during winter.