Use of ANCOVA or transformation

[Bhupinder]

Hi all,

I have three years and the values of dependent variables are quite close for year 2007 and 2008 but very different from 2009 due to weather effects. The effect of treatments is almost similar for individual years but gets diluted when I run years as random. 

Could you please suggest me how to handle this problem? Do I need to use ANCOVA or any transformation? I may see if plant height or any other variable at early stage correlates with dependent variable.

Thanks

Bhupinder