Microtextural and experimental studies have yielded conflicting data on the relative mechanical strengths of muscovite and biotite <Wilson and Bell, 1979; Kronenberg et al., 1990; Mares and Kronenberg, 1993>. We propose a crystal-chemical resolution to this conflict, namely, that (001) dislocation glide in biotite is rate-limited by its fluorine content. Significant F(OH)-1 substitution, and concomitant removal of hydroxyl H+ directed into the interlayer cavity, potentially increases mechanical strength of biotite in two ways: (1) it eliminates K+-H+ repulsion, thereby strengthening the interlayer bonds, and (2) it allows K+ to ''sink'' deeper into the interlayer cavity, the resultant geometry being less favorable to basal slip. In testing this hypothesis we analyzed the naturally deformed biotite studied by Wilson and Bell <1979> and documented its very low F content (XF≤0.02) compared to that of the biotite experimentally deformed by Kronenberg et al. <1990>. Our model and the comparative XF data explain why the biotite of Wilson and Bell <1979> deformed more easily in nature than its coexisting muscovite, whereas the biotite of Kronenberg et al. <1990> was mechanically stronger than muscovite similarly deformed by Mares and Kronenberg <1993>. Our reconciliation of these otherwise conflicting results provides a framework for predicting mechanical strength of natural micas based upon the extent of their F(OH)-1 substitution. Our synthesis highlights the potential role of crystal chemistry in determining mechanical behavior in multicomponent mineral families. Further testing of crystal-chemical effects on rheology will require mineral specimens of both appropriate composition and sufficient size. ¿ American Geophysical Union 1996