The series you cited belongs to so known 'metallic' radius, and it depends on crystal structure of the element, which changes through the row. In short, you cited series, that is not suited for consideration of isolated tendencies.
There are, indeed, several types of atomic radii (covalent with different valur for bonds of different order, van-der-waals radii and radius of cutoff that leaves some amount of electronic density inside the atom). When comparing atomic radii in comparable environment, two main trends are observable: growth of atomic size down the column in periodic table because more electronic shells is packed into same atom, and contraction of atoms towards the end of the row. This is a bit trickier to explain. Essentially, completed inner electronic shell isolate outer shells from nucleus, reducing effective charge of the nucleus that the outer shell 'feels'. Given that, at the beginning of the row outer electrons feels effective charge of 1 around already quite big completed shell, while at the end outer electrons feels effective charge of 8 around compacted inner shell. This is further complicated by 'mixed' status of d-electrons, that are isolated from the nucleus by inner shells much more effectively then p- and especially s- electrons, so they are valence-active in transition elements, but valence-inactive in p-elements.