Material and methods

Five hundred and six two-year-old and mixed-aged ewes of two breeds (Romney and East Friesian Composite (1/2 East Friesian, 1/4 Polled Dorset and 1/4 Texel; EF)) were randomly assigned to either a “Conventional” (CL) or an “Accelerated” (AL) lambing flock. The CL flock consisted of 239 ewes (121 Romney and 118 EF) and the AL flock contained 134 Romney ewes and 133 EF ewes. Thirty-four individual paddocks in a 41.4 ha block were randomly allocated into two blocks of 21.1 and 20.3 ha for the CL and AL flocks, respectively. The experimental period was from March 2003 to August 2006 when lambs from the ewes mated in January 2006 were weaned.

Ewes were shorn in December and in May, regardless of where they were in the reproductive cycle. Lambs were weaned from the ewes and were transport off the property.

Each year, forage crops were fed in summer (hybrid turnip; cv Pasja) and winter (annual ryegrass; cv Hunter) to meet feed demands throughout the year, so that on each block there were approximately 14 ha of permanent ryegrass/white clover pasture and approximately 6 ha of forage crop.

Conventional lamb production flock management Ewe management

condition scores were recorded on the first day of the breeding period (Day P0) and two weeks prior to the first predicted day of lambing (prelamb). Parturition date was recorded for each ewe that lambed.

Lamb management

Within 24 hrs of birth, lambs were weighed, ear-tagged, and dam, litter size and sex were recorded. Lambs were weighed at approximately 35 days of age and at weaning (average age at weaning = 96, 83 and 109 days in 2003, 2004 and 2005, respectively). Weaning occurred at the discretion of the farm manager, which was usually subject to pasture availability, and therefore environmental conditions. To provide a direct comparison of daily weight gain between the CL and AL lambs and to eliminate growth rate differences, a sub sample of CL lambs (n=109) were weighed (unfasted) in 2005, 74 days after the first day of predicted lambing (average age = 66 days).

Accelerated lamb production flock management Experimental design

The accelerated lamb production system was designed to have five breeding periods within each year, beginning 28th March, 9th June, 21st August, 2nd November and 14th January. In order to achieve this, the AL flock was initially divided into three flocks with approximately equal numbers of Romney and EF ewes. Each of the breeding periods were 73 days apart and were 21 days in duration, resulting in a total of 15 breeding periods over the duration of the experiment. Lambing was predicted to begin 146 days after the first day of the breeding period, and lambs were weaned from their dams 73 days after the first predicted day of lambing. The day of weaning

Chapter 3

group, had CIDRs inserted on Day P-11, and at Day P0, joined the group of newly weaned ewes for re-breeding. Ewes were culled if they had three consecutive unsuccessful breeding periods.

Therefore, at the beginning of each 73 day period there was one group of ewes being re-bred (after having weaned lambs or after being diagnosed non pregnant from the previous breeding period), one group beginning to lamb and one group in mid gestation. For example, ewes bred at the 28th March breeding period began lambing 21st August. Lambs from these ewes were weaned and ewes were rebred on 2nd November. Ewes that failed to become pregnant at the March-breeding period were re-bred at the subsequent breeding period (9th June). Ewes mated in June, lambed in November, and were weaned and rebred in January. Non pregnant ewes from June were re-bred in August. This pattern continued for the duration of the experimental period (March 2003 to August 2006 after lambs were weaned from January-mated ewes).

Ewes were fed according to their physiological state (Nicol and Brookes, 2007). Feeding levels were monitored through unfasted liveweight measurements on Day P0, at pregnancy diagnosis (Day P62), two weeks prior to lambing, and at approximately 35 days post lambing. The target for ewes in the AL system was to maintain similar live weights throughout the experimental period.

Ewe management

Ewes were synchronised using progesterone primed controlled internal drug release devices (CIDRs; 0.3 g progesterone; Pharmacia & Upjohn, Auckland, New Zealand). Additionally, equine chorionic gonadotrophin (eCG; Folligon, Intervet Ltd, Auckland, New Zealand) was administered intramuscularly at CIDR withdrawal for the January (400 IU), August (800 IU) and November (800 IU) breeding periods. These

and removed on Day P14.

On Day P62, pregnancy status and the number of foetuses present were determined by transabdominal ultrasonography using a 3.5 Mz transducer. Date of parturition was recorded for each ewe that lambed.

Lamb management

Litter size, lamb sex and dam were recorded for each lamb within 24 hr of birth. Lamb live weight was recorded within 24 hr of birth, at approximately 35 days of age and at weaning (73 days after the first predicted day of lambing).

Statistical analysis

All statistical analysis was done using SAS software (V8, SAS Institute Inc, Cary, NC, 2001). A general linear model (PROC GLM) was used to analyse ewe live weights. The statistical model included lamb production system, year of breeding, breed and ewe age (2-year old vs mixed age). System, breed and year interactions were also tested but were non significant and therefore removed. The model was then re-run without non significant fixed effects and interactions.

Univariate analysis was used to compare the number of pregnant and non pregnant ewes (pregnancy rate). Pregnancy rate was defined as the number of pregnant ewes per ewe exposed to the ram. Pregnancy data were treated as binomial traits, were logit transformed and analysed using a logistical regression model (GENMOD). Values were back-transformed into percentages for presentation. The model for pregnancy rates included lamb production system, year of breeding and breed as fixed effects. Ewe age and ewe live weight were included in the original model but had no significant

Chapter 3

Litter size is defined as the number of lambs in a litter for each ewe that lambed at each lambing period. Litter size at birth and at weaning were analysed as categorical variables with lamb production system, year and breed as fixed effects. Interactions were also tested and removed where there were no significant effects, and the model was then re-run.

The number of lambs born (NLB) and weaned (NLW) per ewe per year, were analysed using the same model. The NLB and NLW are defined as the number of lambs born to each ewe that lambed per year (as opposed to each lambing period). This differs from litter size as ewes in the AL system lambed more than once a year therefore gave birth to and weaned more than one litter within a year.

Lamb live weight at birth and weaning, and average daily weight gain (ADG) were analysed using a general linear model (PROC GLM). The statistical model included lamb production system, year, ewe breed and age (2-tooth vs mixed age), lamb sex, and litter size at birth as fixed effects. Ewe age was not significant in the model so was removed and the statistical model was rerun. Interactions were tested, removed if not significant, and the model re-run with only significant effects.

A general linear model (PROC GLM) was used to compare the growth of CL lambs with the growth rate of AL lambs for a similar period. The statistical model included system, breed, lamb sex and litter size at birth as fixed effects. Birth weight was also tested but had no effect. Ewe age was tested but removed from the statistical model as it had no effect. Interactions were tested and removed where they were not significant. The statistical model was then re-run without ewe age or any interactions.

Lamb mortality was assessed using a univariate analysis which included production system, breed, year and litter size at birth (1, 2 or ≥ 3 lambs) in the statistical

records but no weaning live weight record.