21 - Factors influencing reproductive performance in a range of network situations

W. THORPE, L. COULIBALY, A. DEFLY, G.D.M. d'IETEREN, A. PER ON, G. GRUNDLER, P. HACKER, P. ITTY, J.H.H. MAEHL, K. MAWUENA, G. MORKRAMER, M. MULUNGO, S.M. NAGDA, R. W. PALING, M. PELO, J.M. RARIEYA, A. SCHUETTERLE and J.C.M. TRAIL

Introduction
Materials and methods
Results and discussion
Conclusion
References

Introduction

Reproductive performance is often the major determinant of biological and economic efficiency of livestock production in the tropics. In tsetse-affected areas a factor contributing to poor reproductive performance is said to be trypanosomiasis. Preliminary analyses of parturition intervals of N'Dama cattle and Djallonke sheep recorded during the early stages of the Network research indicated that trypanosome infection post-partum may depress reproductive performance of trypanotolerant livestock (ILCA, 1986). The depressed reproductive performance was not associated with any important liveweight loss caused by trypanosome infection.

Continuing data collection in the Network has provided additional information on which to estimate more accurately any effect of trypanosome infection on parturition interval, the measure of reproductive performance. However, the coefficient of variation of parturition interval, about 35%, is high relative to the coefficients of variation of other health and production traits measured in the Network. This, combined with the smaller number of records available, means that the effect of trypanosome infection on parturition interval will be estimated less accurately than, for example, its effect on PCV traits (d'Ieteren et al., see article 17 of these Proceedings) or on liveweights (Maehl et al., see article 22 of these Proceedings).

Materials and methods

The effects of trypanosome infection and of other factors on parturition interval were evaluated using records for parturitions which occurred between January 1984 and December 1986 at the Network sites of Avetonou, Boundiali and Mushie. For Mushie some intervals, calculated using parturitions occurring in early

1987, were also included. These sites were chosen because of the availability of three years' records for reproductive performance, their relatively high mean trypanosome prevalence and their contrasting combinations of trypanosome species, livestock species, breeds and management systems. Summary descriptions of the Avetonou, Boundiali and Mushie sites are given in Table 1, (see also article 3 of these Proceedings).

Table 1. Summary description of Network sites contributing to the analyses of reproductive performance.

Site	Trypanosome species	Livestock
		Species	Breed	Management system
				Ranch	Village
Avetonou	T. vivax	Cattle	N'Dama	X	X
			Race Locale	X	X
		Sheep	Djallonke		X
Boundiali	T. congolense and T. vivax	Cattle	N'Dama		X
			Baoule		X
			Crossesa		X
		Sheep	Djallonke		X
			Sahelian x
			Djallonke		X
Mushie	T. congolense and T. vivax	Cattle	N'Dama	X

^a Crosses between N'Dama, Baoule and Zebu.

The number of months in which trypanosomes were detected was the measure of trypanosome infection. The periods considered included 3, 6 and 9 months pre-partum and 3, 6 and 8 months postpartum for cattle and 5 months pre-partum and 4 months postpartum for sheep. The pre-partum periods were parts or all of the gestation. The post-partum periods were parts or all of the lactation, up to a nominal weaning at 8 months in cattle and 4 months in sheep.

Analyses were carried out using least-squares fixed-model procedures (Harvey, 1977) including in the model the number of' trypanosome months and other systematic genetic and environmental factors appropriate to the specific Network site, study population and study period. Generally models included the effects, year-season of parturition, herd or flock, dam age, number of trypanosome months, dam liveweight post-partum, dam liveweight change pre-weaning and two factor interactions. Subsequent analyses estimated the relationship between PCV and parturition interval by substituting classes of average PCV (during a specified period) for the number of trypanosome months during the same period. In sheep, the effect of trypanosome infection on litter size was also estimated.

Results and discussion

Sheep

Litter size of sheep

Mean litter sizes at Avetonou and Boundiali were very similar, 1.15 and 1.18, respectively (Table 2). At neither site did trypanosome infection during the gestation have any effect on mean litter size, nor was there any trend suggesting a depression of prolificacy caused by trypanosome infection. Neither was there any relationship between mean litter size and average PCV during gestation.

Table 2. Least-squares means and standard errors for litter size of sheep at trypanosomiasis risk in Avetonou and Boundiali.

Site	No.	Mean	s.e.
Avetonou	227	1.15	0.042
Boundiali	513	1.18	0.046

In Boundiali, Djallonke and Sahelian x Djallonke ewes had the same mean litter size, but litter size was significantly affected by flock, year-season of parturition and dam age. Table 3 presents the results for the effect of dam age, showing that older ewes had significantly larger litters than younger ewes (P<0.001). At these sites some environmental factors, but not trypanosome infection, were important influences on prolificacy of Djallonke sheep.

Table 3. Least-squares means and standard errors for litter size by ewe year of birth classes, Boundiali.

Dam year of birth	No.	Mean	s.e.
1979 and earlier	168	1.27	0.051
1980 and 1981	153	1.19	0.051
1982 and later	192	1.07	0.054
Significance		***

***P<0.001.

Parturition interval

In Boundiali, but not in Avetonou, lambing interval was affected significantly (P<0.05) by trypanosome infection post-partum (Table 4). Trypanosome infection pre-partum did not affect lambing interval at either site. The two-factor interactions between trypanosome infection pre- and post-partum and between trypanosome infection post-partum and other effects did not reach statistical significance, nor were there trends suggesting that interactions might be important.

Table 4. Least-squares means and standard errors for lambing interval (days) of ewes uninfected or infected with trypanosomes during 4 months postpartum, Avetonou and Boundiali.

No. of trypanosome parasitaemias, 0-4 months post-partum	Avetonou			Boundiali
	No.	Mean	s.e.	No.	Mean	s.e.
0	134	248	9	112	228	12
1 or more	18	255	18	69	259	15
Significance		NS			*

*P<0. 05

Post-partum infections were few in Avetonou but frequent in Boundiali where the mean lambing intervals of ewes having none, one, or two or more parasitaemic months post-partum could be estimated for Djallonke ewes (Table 5). Two or more months with detectable trypanosomes did not delay subsequent lambing more than one month with detectable trypanosomes: The independent effects of trypanosome species and number of parasitaemic months could not be estimately accurately as the numbers of ewes with one or two months with detectable parasitaemia caused by T. congolense or by T. vivax were small.

Table 5. Least-squares means and standard errors for lambing intervals (days) of Djallonke ewes with no, one, or two or more trypanosome parasitaemias during 4 months post-partum, Boundiali.

No. of trypanosome parasitaemias 0-4 months post-partum
	No.	Mean	s.e.
0	112	228	13
1	34	265	17
2 or more	35	250	18
Significance		*

*P<0.05

Apart from the effect of year-season of parturition in Boundiali, other systematic environmental effects were not significant sources of variation affecting lambing interval. However in Avetonou, but not in Boundiali, there was a tendency for ewes with heavier post-partum weights and for ewes gaining weight post-partum to have the shortest lambing intervals (P<0.09). There was also a non-significant but consistent trend for ewes with higher average PCVs during the lactation to have shorter lambing intervals (Table 6).

Table 6. Least-squares means and standard errors for lambing interval (days) of classes of average PCV during lactation in Djallonke ewes.

Average PCV during lactation (%)	Boundiali			Average PCV during lactation (%)	Avetonou
	No.	Mean	s.e.		No.	Mean	s.e.
<22	28	267	25
22-24	67	258	17	<24	45	253	12
25-26	69	244	16	24-26	59	249	11
27-29	75	238	15	>26	48	244	13
>29	53	218	42
Significance	NS				NS

In summary, for sheep it can be concluded that neither lambing interval nor litter size of Djallonke ewes was depressed by T. vivax infection in Avetonou, whereas lambing interval but not litter size was depressed by trypanosome infection in Boundiali. The relative effects of T. congolense and T. vivax infections on lambing interval in Boundiali have still to be quantified. There were indications that for Djallonke ewes at trypanosomiasis risk, average PCV post-partum may be related to subsequent lambing interval.

Cattle

Parturition interval

The least-squares means for calving interval varied considerably between sites, with cattle in Avetonou having the shortest calving interval, 424 days and in Mushie the longest, 515 days, with the Boundiali population having an intermediate value 488 days. At each of the three sites cows with trypanosome infection post-partum had longer calving intervals than uninfected cows (Table 7). The difference was some 50 days at each site, but only reached statistical significance (P<0.05) in Avetonou. At none of the sites was there any effect on subsequent calving interval of trypanosome infection pre-partum and no interaction between trypanosome infection pre- and post-partum. The interactions between trypanosome infection post-partum and other effects were not statistically significant.

There were sufficient records at Mushie to investigate further the apparent depressive effect of trypanosome infection by estimating the mean calving intervals of cows which had none, one, or two or more months with detectable trypanosome parasitaemia during the first 8 months post-partum (Table 8). One, but not two or more, parasitaemic months depressed calving interval suggesting that factors other than trypanosome infection were involved in the depression of reproductive performance. Examination of the records of cows having very long calving intervals (if still empty one year after weaning a calf) suggested that these long intervals were not a result of trypanosome infection. These records were therefore deleted from the study population to produce a sample of cows that was more representative of ranch populations maintained under a normal culling regime. The analysis of the culled population showed that under field conditions trypanosome infection post-partum had no effect on the calving interval of N'Dama cows (Table 8).

Table 7. Least-squares means and standard errors for calving intervals (days) of cows uninfected and infected with trypanosomes post-partum in Avetonou, Boundiali and Mushie.

No. of trypanosome parasitaemias 0-6 months post-partum	Avetonou
	No.		Mean		s.e.
0	116		397		20
1 or more	29		452		24
Significance			*
No. of trypanosome parasitaemias 0-8 months post-partum	Boundiali			Mushie
	No.	Mean	s.e.	No.	Mean	s.e.
0	44	465	30	113	489	25
1 or more	38	512	28	75	542	24
Significance	P<0.17			P<0.09

*P<0.05

Table 8. Least-squares means and standard errors for calving interval (days) of N'Dama cows uninfected or infected post-partum with trypanosomes, Mushie.

No. of trypanosome parasitaemias 0-8 months post-partum	No culling			With cullinga
	No.	Mean	s.e.	No.	Mean	s.e.
0	113	492	25	111	481	24
1	36	574	31	30	475	30
2 or more	39	509	30	36	457	26
Significance	P<0.06			P<0.88

^a Original population reduced by retrospective culling of all intervals greater than 799 days, i.e. those cows still empty 1 year after weaning a calf.

The absence of an effect of trypanosome infection postpartum on the calving interval of N'Dama is consistent with the results of the study of Lorenzini et al. (1988) reported in article 18 of these Proceedings. In their experiment, under controlled conditions, N'Dama maintained their reproductive cycle despite being infected with trypanosomes, whereas Boran (zebu) females ceased cycling within 28 days post-infection.

A disadvantage of calving interval as a measure of reproductive performance is that, by definition, only fertile cows are considered. For example in this study only those cows were included which had 2 or more carvings between January 1984 and early 1987. The analysis therefore ignored the possibility that trypanosome infection affected reproductive performance by causing long-term infertility in some cows. However, within the study population, there were few cows, 18 out of 263 (6.8%), which did not calve or calved only once during the study period. These infertile cows (0 or 1 carvings) were compared to the fertile cows (2 or more carvings) for their relative trypanosome infection rates when non-gestating. Both groups had very similar infection rates indicating that there was no association between trypanosome infection and infertility, confirming that trypanosome infection did not depress the fertility of these N'Dama cows receiving good ranch management.

Some other environmental factors did have statistically significant effects on calving interval. For example, year-season of parturition was a significant effect in Avetonou, although not in Boundiali or Mushie. Location of herd and herd-within-location did not affect calving interval in Mushier but, in Avetonou, ranch cows had significantly shorter calving intervals than village cows and cows in "traditional" herds

(those herds whose owners were dependent upon agriculture) had significantly shorter intervals than cows in "modern" herds (those herds whose owners had additional sources of income). In Avetonou there was no effect of the weaning status of the cow on her calving interval, but, surprisingly, those cows in Boundiali and Mushie not rearing a calf to weaning had significantly longer calving intervals than cows rearing a calf. Cows in Avetonou losing weight post-partum had significantly longer subsequent calving intervals than cows maintaining or gaining weight. Cows in Mushie had the same tendency. The effect of liveweight and weight change on calving interval could not be estimated in Boundiali where cows were not weighed regularly until mid-1985.

The effect of the cow's average PCV post-partum on her calving interval was also evaluated. At each of the four sites there was a trend for shorter calving intervals with higher PCV. The effect reached statistical significance (P<0.01) at Mushie where the N'Dama cows are at high trypanosomiasis risk. At Kolo, where trypanosomiasis risk is zero to very low, N'Dama cows with the lowest PCV had the longest intervals and those with the highest PCV had the shortest intervals. PCV may therefore be a useful indicator of performance, regardless of the level of trypanosomiasis risk. More precise estimations of these relationships are required.

Conclusion

It can be concluded from the cattle analyses that the fertility of N'Dama cows in Mushie was not depressed by trypanosome infection post-partum, but that infection may have depressed the fertility of cows in Avetonou and Boundiali. Further investigation of the effect of trypanosome infection and of other factors on cattle fertility is required at these sites and any differential effect of T. congolese and T. vivax infections has still to be quantified.

In addition, within the Network, the study of the effects of trypanosome infection on the reproductive performance of trypanotolerant livestock needs to evaluate possible long-term cumulative effects on dam performance and any effects on male reproduction. The estimation of trypanosome infection x breed and nutrition interactions should also form important components of future Network research. These studies should include the crosses of susceptible breeds with trypanotolerant breeds and, where practical, the susceptible breeds.

To achieve these objectives livestock populations are required which are at high trypanosomiasis risk and which will therefore provide large numbers of animals infected naturally with trypanosomes. These naturally infected animals can then be compared with contemporary controls.

References

ILCA. 1986. The African Trypanotolerant Livestock Network: Indications from results, 1983 - 1985. ILCA, Addis Ababa.

Harvey, W.R. 1977. User's Guide for Least-Squares and Maximum Likelihood Computer Program. Columbus: Ohio State University.