The problem as to why alpha-lactalbumin, in the absence of Ca(2+), forms a molten globule intermediate, in contrast to its structural homologue lysozyme, has been addressed by the construction of chimeras of human lysozyme in which either the Ca(2+)-binding loop or a part of helix C of bovine alpha-lactalbumin were transplanted. Previously, we have shown that the introduction of both structural elements together in the lysozyme matrix causes the apo form of the resulting chimera to display molten globule behavior during the course of thermal denaturation. In this article, we demonstrate that this molten globule character is not correlated with the Ca(2+)-binding loop. Also, the Del 101 mutant in which Arg101 was deleted to simulate the alpha-lactalbumin conformation of the connecting loop between helix C and helix D, does not show a stable equilibrium intermediate. Rather, the molten globule character of the chimeras has to be related with a specific part of helix C. More particularly, attention is drawn to the four hydrophobic side-chains I93, V96, I99, and L100, the lysozyme counterparts of which are constituted of less bulky valines and alanine. Our observations are discussed in terms of decreased stability of the native form and increased stability of the intermediate molten globule.