The most likely experimental design used to provide evidence for the success in blood fractioning is:
**D. Blood donations were collected and separated by their different densities in the mixture.**
This method aligns with common techniques in blood fractionation, where components of blood (such as plasma, platelets, and red blood cells) are separated based on their densities, typically using centrifugation. The other options presented (separation by size, boiling point, or magnetic attraction) are not standard methods for blood fractionation. Therefore, option D is the most appropriate choice.