The correct option is A G1G2
Let x1,x2,..xn be n obeservation corresponding to variate x and y1,y2,..yn be n observation corresponding to variate y
Then by defination
G1=(x1.x2...xn)1n and G2=(y1.y2...yn)1n
Now the geometric mean of this is given by
G=(x1y1.x2y2.x3y3...xnyn)1n=(x1.x2...xny1.y2...yn)1n=G1G2