AbstractIt has been impossible, up to now, to focus ferritin (a large, ca. 450 000 Da, ironstorage protein ubiquitous in mammalian systems) in immobilized pH gradients (IPG). This was not a sieving problem, as ferritin focuses readily in a 4%T polyacrylamide (or polytrisacryl) gel in conventional carrier ampholyte isoelectric focusing. A strong interaction with the IPG matrix was suspected, but an auto‐buffered (pH 6.0) soluble IPG polymer failed to precipitate ferritin from solution, although it altered its visible spectrum, suggesting the formation of soluble complexes. By a Takeo plot, the dissociation constant (Kd)of the ferritin‐IPG matrix complex was estimated to be 0.1 mM. However, even a 100‐fold dilution of Immobilines into the IPG matrix with a concomitant addition of competing ions (up to 3 % carrier ampholytes) failed to produce good focusing patterns. An unsuspected, hitherto un‐reported, precipitating power of free Immobilines in solution on the ferritin macroion was found to be the key to the riddle. The precipitating ions are (in order of aggregation power): pK 9.3 > pK 7.0 > pK 6.2. None of the acidic Immobilines (pKs 3.6, 4.4 and 4.6) and, among the basic, only the pK 8.5, exhibit any precipitation power on ferritin and other complex protein mixtures. In a new IPG pH 4–5 interval, containing only the pK 4.6 as buffering and the pK 8.5 Immobilines as titrant ions, excellent focusing patterns of ferritin were obtained. It is hypothesized that the precipitating Immobilines act as cationic detergents, and are able to cross‐link macroions with a mixed‐type, hydrophobic‐ionic interaction at the two extremities of the molecule. Guidelines are given for handling this major problem on a short‐term basis, the longterm goal being the replacement of the ‘unfavorable’ Immobilines (pKs 9.3, 7.0 and 6.2) by new chemical structures of the desirable type (the non‐precipitating pK 8.5 species).