The California Mastitis Test has previously been adapted for use in an inline, cow-side sensor and relies on the fact that the viscosity of the gel formed during the test is proportional to the somatic cell concentration. In this paper, the use of capillary and rotational viscometry was compared in light of the expected rheology of the gel formed during the test. It was found that the gel is non-Newtonian, but the initial phase of viscosity increase was not due to shear dependence, but rather due to the gelation reaction. The maximum apparent viscosity of the gel was shear dependent while the time it took to reach the maximum was not truly shear dependent, but was rather dependent on the degree of mixing during gelation. This was confirmed by introducing a delay time prior to viscosity measurement, in both capillary and rotational viscometry. It was found that by mixing the reagent and infected milk, then delaying viscosity measurement for 30 s, shortened the time it took to reach maximum viscosity by more than 60 s. The maximum apparent viscosity, however, was unaffected. It was found that capillary viscometry worked well to correlate relative viscosity with somatic cell count, but that it was sensitive to the reagent concentration. It can therefore be deduced that the rheology of the gel is complicated not only by it being non-Newtonian, but also by the strong dependence on test conditions. These make designing a successful sensor much more challenging.