Tuberous sclerosis: CT and MR findings.

solubility and means of purification , given its yet imperfect structural and compositional definition. The stain reaction requires the presence of only two identical or similar metal ions in different oxidation states to generate the intense blue color. For example, colorless solutions of Fe3+(CIO.)J and K. Fe+(CN)6 can be mixed to produce the bright blue color. The ferritin stain uses an antibody specific for the well-characterized iron-protein complex that binds iron for storage and transport. It is clear from our results [1] that the distribution of ferritin immunoreactivity and Perls positivity are different. Ferritin immunoreactivity has a wider distribution than Perls ' positivity does. This difference in distribution cannot be explained by the differences in solubility as implied by Bizzi et al. Differences in distribution may represent differences in reaction rates for different forms of iron. Calibrating response curves for tissue sections by using polychromatic diffuse light to obtain quantitative measurements of iron concentration is complex and was well beyond the requirements of our study, as discussed elsewhere (Thulborn et al. , poster presented at the annual meeting of the Society of Magnetic Resonance in Medicine, 1989). It was not intended to imply that ferritin does not react with the Perls stain but rather to show that the specific protein ferritin is present in chronic hemorrhage. None of the assumptions implied in the letter of Bizzi et al. were made or were necessary for this work. Others [2] have suggested that a positive Perls stain should not be used to imply that ferritin only is present. The brain contains a sizable pool of iron that is not characterized biochemically and that likely reacts with this stain. Animal models of iron metabolism will be an important component of detailed biochemical analysis of all iron pools. I thank Bizzi et al. for their support for the conclusions of our paper, if not the semantics.