I agree with Mark. Check out IBM Bodoni.

And here's the apostrophe.

I was concerned about those two things as well, but since the sample is so rough, it's possible that the k connection and the t curve simply filled in in the sample. Hope springs eternal.