The mechanics of shear failure is a key to understanding a wide range of geological processes, from landslides on the surface to faulting in the Earth's interior. It is now a well-established fact that ductile materials yield under a threshold stress condition, forming shear bands of finite length, which eventually influence the paths of their bulk failure. Compression test experiments show mechanically homogeneous isotropic solids undergo shear failure to produce a conjugate set of shear bands or fractures symmetrically oriented to the applied compression direction (Anand & Spitzig, 1980;Bowden & Raha, 1970;Hutchinson & Tvergaard, 1981;Tvergaard et al., 1981). Employing various yield criteria, many workers have provided theoretical solutions to predict the band orientation as a function of material parameters, such as coefficient of internal friction, dilatancy factor, and strain hardening and softening properties (Anand & Spitzig,

Ductile yielding of rocks and similar solids localize shear zones, which often show complex internal structures due to the networking of their secondary shear bands. Combining observations from naturally deformed rocks and numerical modelling, this study addresses the following crucial question: What dictates the internal shear bands to network during the evolution of an initially homogeneous ductile shear zone? Natural shear zones, observed in the Chotonagpur Granite Gneiss Complex of the Precambrian craton of Eastern India, show characteristic patterns of their internal shear band structures, classified broadly into three categories: type I (network of antithetic low-angle Riedel (R) and synthetic P-bands), type II (network of shear-parallel C and P/R bands) and type III (distributed shear domains containing isolated undeformed masses). Considering strain-softening rheology, our two-dimensional viscoplastic models reproduce these three types, allowing us to predict the condition of shear band growth with a specific network pattern as a function of the following parameters: normalized shear zone thickness, bulk shear rate and bulk viscosity. This study suggests that complex anastomosing shear-band structures can evolve under simple shear kinematics in the absence of any pure shear component.

<p>Development of brittle and brittle-ductile shear zones involve partitioning of large shear strains in bands, called C-shear bands (C-SB) nearly parallel to the shear zone boundaries. Our present work aims to provide a comprehensive understanding of the rheological factors in controlling such SB growth in meter scale natural brittle- ductile shear zones observed in in Singbhum and Chotonagpur mobile belts.  The shear zones show C- SB at an angle of 0°- 5° with the shear zone boundary. We used analogue models, based on Coulomb and Viscoplastic rheology to reproduce them in experimental conditions.</p><p>These models produce dominantly Riedel (R) shear bands. We show a transition from R-shearing in conjugate to single sets at angles of ~15<sup>o</sup> by changing model materials. However, none of the analogue models produced C-SB, as observed in the field. To reconcile the experimental and field findings, numeral models have been used to better constrain the geometrical and rheological parameters. We simulate model shear zones replicating those observed in the field, which display two distinct zones: drag zone where the viscous strains dominate  and the core zone, where both viscous and plastic strains come into play.  Numerical model results suggest the formation of  C- SB for a specific rheological condition. We also show varying shear band patterns as a function of the thickness ratio between drag and core zones.</p>