Effects of concentrate feed levels and herd management on roughage feeding behaviour of dairy goats

• About 10,000 years BC, goats were the first ruminants to be domesticated. They were appreciated for their small body size, resistance towards diseases and undemanding food requirements. To date, most goats are being kept in developing countries as "poor man's cows". Due to marketing of goats' cheese as well as chevon (goat meat) as a new delicacy, however, the number of goats in highly industrialized Germany, for instance, has also increased in recent years. • In Europe, goats, like cattle, are bred for high performance (dairy breeds, meat breeds). In order to maximize genetically predisposed performance, animals are fed concentrate supplements (e.g., wheat, soy, maize). This is considered problematic in the global context, with complex discussions concerning destruction of rain forests, climate change and world food supply. • Goats are very undemanding in their food requirements, but display a pronounced ability for feed selection if the opportunity arises. Moreover, their facultative bipedalism and well-developed climbing skills enable goats to utilize leaves and young twigs as additional feed sources. Concerning the digestion of cellulose, the goat's digestive system is even more efficient than are those of sheep or cattle. • Given the specifics of goat nutritional physiology as well as feed selection behaviour, these animals were predestined to feature in the thesis presented. The aim of this dissertation was to investigate potential compensatory patterns of roughage feeding behaviour at two different stages in the life of dairy goats kept under commercial husbandry conditions (breed: "German improved fawn" = "Bunte Deutsche Edelziege"). For this, behaviour of 16 young dairy goats (age: approx. 8 months; of which 9 goats were genetically hornless) as well as 51 and 41 adult horned dairy goats (age: 2 to 7 years; with difference in herd size due to different seasons), respectively, was analyzed within different study approaches. Analyses were effected at individual level as well as at group/herd level. Study animals were kept under organic standards (EU-organic-regulation No 889/2008 with code of practice No 834/2007; Bio Suisse guidelines). They spent half the day at pasture, for the rest of the day and during the night they were stabled in a loose-housing barn with ad lib. hay at the feeding table. All observations were conducted under commercial husbandry conditions and incorporated into the daily routine at the research station of the Johann Heinrich von Thünen-Institute of Organic Farming (Wulmenau, Schleswig-Holstein). • The three studies comprising the thesis presented here, cover the full spectrum of ethological approaches, i.e., methodological, experimental and empiric. • Data collection for the methodological study included direct observations (behaviour sampling), digital photography (instantaneous scan sampling) and digital video recordings (continuous recording followed by instantaneous time sampling). Data for the experimental study were collected using direct observations (behaviour sampling) and digital video recordings (continuous recording followed by instantaneous time sampling). With respect to the empirical study, data were collected exclusively by direct observation (continuous focal animal sampling). Data were analyzed using Microsoft© Excel and statistical software "R" (Version 3.2.1). • The preceding methodological study served to assess reliability and efficiency of different visual methods for recording and evaluating feeding behaviour exhibited in the barn and at pasture. Behaviour was recorded simultaneously using direct observation, still photography and videos at 3-min intervals over a total of 6 hours (video footage was continuous and subsequently broken down into 3-min intervals). In a first step, agreement between measurements of the three sets of data with a reference data set was quantified using the Bland-Altman method. To date, this statistical method for assessing agreement between two methods of measurement is rarely applied in ethological research; it is predominantly used in clinical research as the standard test. (In medical studies, appropriateness of the test being used for comparing established and alternative treatment methods is of literally vital importance, as human lives depend on it.) Reliability assessment was based on the "median of percental deviation", which had been specifically devised in the course of the methodological study and served as a quantitative cut-off criterion. The reliability assessment conducted in this investigation is universally applicable. The procedure developed hence has the potential to contribute towards increasing the degree of comparability between future studies. In a second step, data sets corresponding to interval lengths of between 3 and 12 min were generated and compared with a data set that had been evaluated at 1-min intervals. • Only continuous video footage proved appropriate for recording the feeding behaviour of larger herds (n = 41 and 51 individuals, respectively) in the loose-house and at pasture, since continuous videos allow to identify inconclusive behaviour (and hence complete the data set) by winding the recording back and forth. For data collected in the barn, the longest interval yielding representative data was 3 min. At pasture, feeding frequencies were five times as high as in the barn, which additionally rendered 5- and 8-min intervals sufficiently reliable. The temporal organization of the respective behaviour (frequency, duration) thus determined the selection of the longest, sufficiently reliable interval. • Three years prior to data collection, the 41 adult goats had been assigned to two treatment groups, which received different concentrate rations, though they continued to live as one herd. Concentrate for Group A (n = 18) equalled 40 % of total yearly amount of dry matter (= 1000 g per animal and day during lactation), Group B (n = 23) received 10 % concentrate (= 100 g). Concentrate was fed during milking. In this experimental study, roughage feeding behaviour (hay, straw, grass) in the barn and at pasture was recorded on video over a total of 30 hours. Footage was subsequently evaluated at 3-min intervals (resulting in 600 data points per animal = 24.600 in total). Behaviour sampling was used to record dominance-related interactions during a total of 97 hours (i = 4.017). Individual 240 day milk yields, live weight and milk fat-to-protein ratio were included in the analyses. Results supported a positive relation between concentrate level and productivity. In both groups, fat-to-protein ratio indicated a balanced nutritional state. Group B animals exhibited lower weight ( 6 %) as well as milk yield ( 17 %) than those of Group A. Moreover, frequencies of roughage feeding were significantly higher for Group B animals than for Group A goats. In conjunction, these findings suggest roughage feeding to be at least partially compensatory. Dominance indices were positively correlated with hay feeding frequencies in the barn, and significantly negatively correlated with grazing at pasture. For low-ranking goats, the compensatory time at pasture is of particular relevance. High-ranking animals are far less able to monopolize roughage at pasture (i.e., grass) than in the barn (hay, cut grass). In the latter setting, low-ranking animals are often kept from roughage feeding. The losses incurred through lower animal productivity due to reduced amounts of concentrate fed should be weighed against the saved costs of concentrate feed, as there might well be a financial net gain. In terms of ethical considerations concerning animal health, care should be taken to keep concentrate feed levels to the metabolic minimum required in order to avoid severe diseases (e.g., rumen acidosis). • In the empiric study, 16 young female goats were assigned to two groups (following common husbandry practice, allocation was based on weight) and introduced into the established adult herd (n = 41) in two batches. Approx. two months older and also heavier, Batch 1 goats joined the herd three weeks earlier than Batch 2 goats. Data were collected in two observation periods separated by a two-week gap. Roughage feeding behaviour in the barn was investigated on a second-by-second basis for a total of 174 h. During data recording, dry-roughage feeding behaviour was differentiated into four categories of increasing degree of spatial competition avoidance ("feeding normally", "feeding vigilant", "feeding through the barrier" and “feeding straw”). Individuals varied considerably with respect to combinations of the four category proportions ("profiles"). By arranging the profiles along two dimensions (quantity and quality) of their predominant category, five major types and a total of nine subtypes could be differentiated. In each of the two observation periods, five of the nine types had been realized; these were, however, not necessarily identical - nor was the allocation of individuals to the types. Without exception, the "Mixed Type" contained a minimum of three categories, with each of them amounting to < 50 %. "Primary Types" were defined by one of the categories equalling or exceeding 50%. They were sub-divided into "pronounced" (≥ 75 %) and "moderate" forms (≥ 50 % to < 75 %). Newly introduced and hence unfamiliar animals need to find their place in the herd's social structures (e.g. bonds and dominance relations). The first observation period was characterized by a marked preponderance of the behavioural category associated with the highest degree of spatial competition avoidance. Prevalence of this category decreased from observation period 1 to 2; feeding normally, however, remained the exception throughout the entire duration of the study. The 16 young goats experienced a double disadvantage with respect to dominance-related access to the feeding table (due to being not only "new kids on the block" but also the youngest individuals). Analyses of the young goats' feeding types suggested that they developed individual behavioural patterns. These patterns allowed them to avoid confrontations with higher-ranking (adult) goats at the feeding table and enabled them to ingest roughage despite their low competitive abilities. • The two studies on compensatory roughage feeding behaviour raise the importance of roughage (grass, hay and even straw) vis-à-vis concentrate feed. In feeding management, competition avoidance behaviour of low-ranking herd members may be utilized to maximize overall herd productivity. Comprehensive and detailed knowledge of roughage feeding behaviour including aspects of competition compensation, thus benefits animals and livestock owners alike by increasing animal welfare as well as economic efficiency.