The study herds and flocks
Field methods
Data analysis
Additional sources of information
Following a preliminary survey in early 1978, herds and flocks in six locations in the millet subsystem and in four locations in the rice subsystem were selected for study. These units were then followed on a regular basis until June 1984. At the start of the survey there were approximately 600 head of cattle in the herds and some 700 head of goats and sheep.
At the end of June 1984, when the analysis was undertaken, there were records for over 1150 cattle and 5000 small ruminants, both these totals including those animals that were present at the beginning of the study. Most of the animals entering the herds did so by birth, although there was a small number of purchases and some animals were brought in as gifts, loans or exchanges.
At the start of the survey, all cattle were individually identified by means of numbered plastic or metal ear tags. Details of the animal's age (determined by dentition and/or by owner questioning), its function in the herd and its reproductive history (if a female) were obtained. From these data, an individual record card was established. All animals subsequently entering the herds were treated in the same way, and individual record cards were also created for these animals.
Visits were made to cattle herds on a regular basis, usually at intervals of about 4 to 6 weeks but more often during the main calving period of late April to early August. Owners cooperated in defining the events - births, deaths, purchases and sales - which had taken place during the time period which had elapsed since the previous visit. Details of all these events were entered on the individual record cards or new record cards were created, as appropriate.
Goat and sheep flocks were visited at intervals of about 15 days. Young animals were weighed on most of these occasions, while older animals were weighed on every second visit.
Lighter animals (estimated as less than 100 kg) were weighed on most of the regular visits by suspending them in a sling attached to a dial-type spring balance. Heavier animals were weighed in a proprietary brand, transportable weigh crush. Weighings of heavier cattle were made at less frequent intervals, usually four times each year: these were timed to coincide with the end of the rains (September/October), the beginning of the dry period (late November to early January), the early part of the hot dry period (March/April) and the end of the dry season and beginning of the rains (July). On occasions, logistical and administrative problems made it impossible to adhere to this timetable. Some weighings of adult cattle therefore appear in the records for other months of the year.
The individual records for each animal were entered on a computer and the following subsets were established from the main data base, relating to the animal's identity and aspects of its productivity:
· Identity: individual number, system of management, herd identity, reason for entry, sex, parity, type of birth, date of entry, reason for exit, dam number.· Reproductive performance: age at first parturition, date of parturitions, interval between successive parturitions, identity of young, sex of young, the parity (order in its dam's reproductive record) of young. Where previous reproductive history of cows of mature age was unknown, the first recorded parturition was set arbitrarily at three, and where this parameter was unknown for goats or sheep, the first recorded parturition was set at four.
· Weight: two subsets were established; one for animals of known birth date from which it was possible to calculate growth rates and the effect of some variables on it, and one for animals of unknown birth date which was used primarily for establishing weights for age in the older age groups and for determining the magnitude of seasonal weight changes in these animals.
· Dentition pattern: for a small number of animals this was used to determine ages at eruption of permanent incisors, but the main use of this subset was in the weight-for-age data at older ages.
· Milk production: small sets of records were established on milk offtake for human consumption for a few animals in one herd of cattle and for the total milk production of some Macina sheep.
Usual compilation and statistical techniques using a pocket calculator were used for many preliminary analyses and as an aid to verification and preparation of data. Most final analyses were carried out using recognised software packages, particularly SPSS, BMDP, GLIM and the Harvey (1977) model for the generalised least-squares procedure suitable for use on data with unequal subclass numbers. The estimated least-squares means generated by this analysis, being adjusted for the unequal subclass numbers, may differ from the observed means.
Effects in the Harvey model usually included the fixed effects of:
- subsystem: millet, rice- month of parturition or birth
- parity: 1, 2....n end '9' when unknown but considered to be ³ 3 for cattle and ³ 4 for small ruminants:
- type of birth: single, twin, triplet or multiple where twin and triplet births were combined
- sex: female, male
- herd within system
- year, and
- interactions of some of the above main effects.
When adequate data were available, the random effects of dam within flock or within system were used in a mixed model.
The residual mean square was used as the error term in the Harvey model to test the significance of all differences. Linear contrasts of least-squares means were computed to determine the significance of differences between groups.
For cattle, a total of 519 births from 274 cows or heifers and 244 calving intervals were recorded during the study. The weights of animals of known birth date totalled almost 6000 records: all weights, including those from mature animals, totalled over 10 000 records.
In small ruminants, 3605 parturitions gave rise to 4049 young and more than 2000 parturition intervals. Almost 40 000 weights were recorded.
Considerable additional information was obtained from primary or secondary sources and has been incorporated in this report. In particular, this information related to the production systems (Section 3), to the general field of management (Section 5), to herd demography (Section 6) and to some aspects of reproductive performance (Sections 7 and 11) and productivity (Sections 10 and 14).
This information was obtained by means of:
· specific short-term studies,· direct observations in the field,
· structured and unstructured interviews with owners and herders,
· data collected by students of the Rural Polytechnic Institute, supervised by ILCA staff and published as dissertations,
· information extracted from internal reports (Programme Documents) of the Arid and Semiarid Zones Programme, and
· other relevant literature.
These sources have been acknowledged in the text and listed in Part 5 of this report.
