*3.6. Behavioural Measures*

Behavioural outcomes were widely reported and ranged from evaluation of spontaneous behaviour, such as eating (ascertained through bodyweight), vocalisation or inactivity, through to the use of well-established behavioural tests of anxiety or a ffective state [2,8,12,30,32].

Bodyweight loss can result from a variety of causes all of relevance to well-being. These include disease, poor or lack of nutrition, as well as eating behaviour which is potentially compromised by a stress response or trauma associated with a procedure. However, stable bodyweight does not necessarily imply that well-being is not impaired, or even positive in nature [53]. Whilst the use of bodyweight alone as a measure of welfare is fairly crude and non-specific, it is a commonly used surrogate parameter for welfare [54].

Nest building is a spontaneous behaviour that has been proposed to represent a 'luxury' behaviour which is highly motivated but non-essential in the laboratory [55]. As such these behaviours are generally the first to be reduced during times of stress [44]. Therefore, poor or reduced performance in this behaviour may indicate a reduction in well-being [55].

Elevated plus maze is used as an assay for anxiety-related behaviour, and typically utilises several di fferent outcome measures to ascertain level of anxiety behaviour, with the general presumption being that increased open arm activity is anti-anxiety behaviour [56]. Open arm activity can be measured as the number of entries or duration. A range of other parameters are frequently collected in this test, including velocity in open and closed arms and distance covered. These are, however, typically a measure of locomotor activity rather than anxiety per se [57].

The open field test is used to gather information on ambulation and emotionality [58]. Ambulation or activity can be measured using total distance travelled in the test. Thigmotaxis is used as a measure of anxiogenic behaviour, with thigmotaxis increasing as anxiety increases. This is typically measured through entries into the central zone or time spent in the centre versus the periphery [58].

### *3.7. Blood Sample Quality Measures*

Based on our restricted definition of measures of blood sample quality, measures of hemolysis and clotting were considered in our synthesis. Hemolysis is the most common pre-analytical sources of error in clinical laboratory and generally leads to sample rejection and the need for blood re-draw. The finding therefore has animal welfare as well as experimental implications. Furthermore, invisible hemolysis can lead to discharge of cell constituents and false results [59]. Clotting may occur where the blood is slow to fill the collection container, or when considerable manipulation of the vein by the needle has occurred. The presence of clot can therefore give a good indication of the ease with which sample can be collected via the particular route but may also be lessened by operator experience [60]. These samples are unable to be analysed for most laboratory tests.
