So, in VRML we use a 3 coordinate system; X, Y, & Z. X is to the left and right, Z is towards or away from you, and Y is up or down. In order to get something to stack up, you would use the same X and Z, and just change the Y values. Thusly.....

#VRML V2.0 utf8
Transform {
    translation  0 0 0  ### set at location 0,0,0
    children
    [
        Box { size 1 1 1 }
    ]
}
Transform {
    translation 0 1 0  ### move this ball up one meter
    children
    [
    Sphere { radius 0.5 }
    ]
}
Transform {
    translation 0 2 0  ###move this cone up 2 meters
    children
    [
        Cone {
            bottomRadius 0.5
            height 1
            side TRUE
            bottom TRUE
        }
    ]
}

 

http://philliphansel.org/simple_translation.wrl