Genetics of resistance to bacterial and parasitic infection / edited by D. Wakelin and J.M. Blackwell.

Date:
1988
Catalogue details

Licence: Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0)

Credit: Genetics of resistance to bacterial and parasitic infection / edited by D. Wakelin and J.M. Blackwell. Source: Wellcome Collection.

    61/308 page 47
    Mode of inheritance of host response 47 Table 2.8. Angle of per cent fluid consumed as the specified solution in a two-bottle choice test. ' Angle is arctan [sq rt (per cent/100)] in degrees. Each box contained three mice of the specified strain and the fluid consumed was recorded over 2 days, with the position of the bottles being switched each day. Data was collected on box 1 first, then the whole experiment was repeated with box 2. (Unpublished data of Fawdington and Festing). depend on the experimental design. In many cases a completely randomized design will have been used in which samples of animals will have been tested more or less at random over a period of time. In this case, a one-way analysis of variance would be appropriate. In the illustrative data, however, box 1 of all 16 strains was tested first, and then after an interval the whole experiment was repeated. This represents a randomized block experiment design (Cochran and Cox 1957; Cox 1958), and must be analysed using a two-way analysis of variance. The analysis of variance is shown in table 2.9. An F-test is used to test the overall significance of differences between strains. Such differences are usually so obvious as not to need a formal test of significance. However, the error mean square, which is an estimate of the pooled within-group variance, is needed
    digitised image 1
    digitised image 2
    digitised image 3