Normal and non-parametric repeated test data

Hi All

I have done some experimental work and collected data for the conductivity of water with various concentrations of salt. The experiment collected data every 15 mins for 24 hours. Some of the data produced is Normal and other data sets are not normal. Hence I don't know what stats I should do on this data.

What I have collected and analysed is the following

Tests of Normality
Kolmogorov-Smirnova Shapiro-Wilk
Statistic df Sig. Statistic df Sig.
Water .231 86 .000 .825 86 .000
0.1% NaCl .071 86 .200 .987 86 .528
1% NaCl .052 86 .200 .991 86 .844
5% NaCl .229 86 .000 .706 86 .000

What I want to do is a pairwise comparison between these data sets to see if the changes between them is significant.

I have been told that a Kruskal-Wallis Anova will do what I want, but given the data is recorded over time I am unsure. Can anyone help me decide what test I should be doing?