Context :

Aims : We study the impact of a subsolar metallicity on various properties of non-rotating and rotating stars , such as surface velocities and abundances , lifetimes , evolutionary tracks , and evolutionary scenarios .

Methods : We provide a grid of single star models covering a mass range of 0.8 to 120 M _ { \sun } with an initial metallicity Z = 0.002 with and without rotation .
We discuss the impact of a change in the metallicity by comparing the current tracks with models computed with exactly the same physical ingredients but with a metallicity Z = 0.014 ( solar ) .

Results : We show that the width of the main-sequence ( MS ) band in the upper part of the Hertzsprung-Russell diagram ( HRD ) , for luminosity above \log \left ( L / L _ { \sun } \right ) > 5.5 , is very sensitive to rotational mixing .
Strong mixing significantly reduces the MS width .
Here for the first time over the whole mass range , we confirm that surface enrichments are stronger at low metallicity provided that comparisons are made for equivalent initial mass , rotation , and evolutionary stage .
We show that the enhancement factor due to a lowering of the metallicity ( all other factors kept constant ) increases when the initial mass decreases .
Present models predict an upper luminosity for the red supergiants ( RSG ) of \log \left ( L / L _ { \sun } \right ) around 5.5 at Z = 0.002 in agreement with the observed upper limit of RSG in the Small Magellanic Cloud .
We show that models using shear diffusion coefficient , which is calibrated to reproduce the surface enrichments observed for MS B-type stars at Z = 0.014 , can also reproduce the stronger enrichments observed at low metallicity .
In the framework of the present models , we discuss the factors governing the timescale of the first crossing of the Hertzsprung gap after the MS phase .
We show that any process favouring a deep localisation of the H-burning shell ( steep gradient at the border of the H-burning convective core , low CNO content ) , and/or the low opacity of the H-rich envelope favour a blue position in the HRD for the whole , or at least a significant fraction , of the core He-burning phase .

Conclusions :