Jacobian matrix evaluation doubt!

Is there a way to evaluate a jacobian matrix having symbolic variables that change with every time step? Here is the code that I am running:

syms yA yG yC

k1 = 1;

k2 = 1;

k3 = 1;

k0 = k1 + k3;

f = [(-k1-k3)*(yA^(2)); (k1*(yA^(2))) - (k2*yG); (k3*(yA^(2))) + (k2*yG)];

v = [yA yG yC];

J = jacobian(f,v);

t = 0:0.1:10;

yA = 1./(1 + (k0*t));

eta = k2./(k0*yA);

a = (k1*k2)/(k0^(2));

yG = a*exp(-eta).*(((k0/k2)*exp(k2/k0)) - (exp(eta)./eta) + mfun('Ei',eta) - mfun('Ei',k2/k0));

yC = 1-yG-yA;

for i = 1:length(t)

    j = J([yA(i) yG(i) yC(i)]);

    %j = subs(J, [yA yG yC], [1 0 0]);

    lambda = eig(j);

    r = double(lambda); %converts a symbolic matrix to a matlab numeric form

    b(i) = max(abs(r));

    c(i) = min(abs(r));

    SR(i) = b(i)./c(i);

end

This line: j = J([yA(i) yG(i) yC(i)]); is not being evaluated properly? Am I wrong somewhere?

Thank you for the help!

Regards

Ushnik