Okay, for question 2) (I'm not completely sure if I'm right but I'm just going to harvest a guess...)
So reliability judges the consistency of the results of an experiment. This means that if an experiment is reliable, it will always obtain the same results when performed with other experimenters, other participants, etc.
However, validity judges how accurately the experiment is testing the cause-effect relationship. Thus, if an experiment shows validity, then it is accurately testing for a relationship between the independent and the dependent variables.
So, reliability can be possible without validity because an experiment can consistently produce the same results when performed with other experimenters and other participants, but it doesn't have to accurately test the cause-effect relationship.
But for an experiment to be valid, and thus accurately testing the relationship, it wouldn't make sense if it didn't consistently test this relationship. For as the experiment is ACCURATE, why would it obtain different results with another experimenter if performed accurately? Thus validity requires reliability.
Sorry if that doesn't make sense - I know I need to use an example but I'm in my pjs and need to go have breakfast.
I think that's the answer anyway (but not 100% sure). I would answer 1) as well but like I said, need my breakie.