Using GPower3 to calculate sample size

I ran the following calculations in GPower3 but I am doubting my method and I am looking for suggestions but also perhaps a beginners way of checking this mathematically.

I have a one way test, with an r of .25 an alpha of .05 and a power of .80, I came up with a required sample size of 95 which is at least close to one of the multiple choice answers given.
However I have come unstuck with a second question where I need again to calculate a sample size useing GPower3.

I have a one way hypothesis again, an effect size Cohens D of .30, my t test has an equal number of male/female participants, a target power of .80 and an alpha of .05
- my answer this time comes no where near the range of 580-278 sample size offered in the multiple choice.

I have been trying hard on this, but I am truly stuck as standard deviation seems to come into the equation somewhere but I can't see how I would compute that with no data... I hope to learn a good deal more on stats in time...