At typical imaging fields, the transverse relaxation rates 1/T2 of the protons of soft tissue are much greater than their longitudinal rates 1/T1. Because of this, clinical magnetic resonance images are generally collected using relatively short values of TR, an approach that both increases comfort for the patient and reduces medical costs. As a result, image contrast is dominated by the 1/T2 values of the tissue protons. Currently, small single-ion paramagnetic complexes - Gd-DTPA is the prime example - are being used to enhance contrast in clinical magnetic resonance imaging (MRI). However, such agents contribute comparably to 1/T1 and 1/T2 so that their utility is greatest when introduced into body fluids, for which 1/T1 and 1/T2 are also comparable; they are much less useful for enhancing contrast of soft tissue. For this, one must look elsewhere, to rather large aggregates of paramagnetic ions, which may either be paramagnetic or ferromagnetic. Iron in its many chemical and biochemical forms, both exogenous and endogenous, is important in this respect. Its presence in ferritin and hemosiderin - in excess in some diseases - is one example; deoxyhemoglobin in cells and methemoglobin in blood pools from trauma are others in which endogenous iron in several oxidation states is important. Magnetic particulates of various iron oxides are now being used as exogenous agents for enhancing 1/T2 preferentially at imaging fields. Predicting contrast enhancement under such circumstances can become rather complex, not because the theory is difficult, but because the underlying concepts are subtle. We recently reexamined the theory of 1/T2 as it applies to situations relevant to preferential enhancement of 1/T2 of tissue protons; here we present the phenomenology, based on this theoretical work, of the behavior of both 1/T1 and 1/T2 - at all values of magnetic field - for solutions of Mn2, complexes and ferritin, and suspensions of erythrocytes, all treated initially as magnetized spheres.