For describing induced dipole I have usually seen the following equation describing the situation, 
$$P_{i}=\alpha_{i,j}E_{j}+\frac{1}{2}\beta_{i,j,k}E_{j}E_{k}$$
where $P_{i}$ is the $i^{th}$ component of the induced dipole moment, $\alpha$ is the polarizability and $\beta$ is the hyper-polarizability.

However, recently I have been reading a paper where I see quadrupole coupled terms coming in, like dipole-quadrupole polarizability , for example as shown below;
$$P_{i}=\alpha_{i,j}E_{j}+\frac{1}{3}A_{i,jk}\frac{1}{2}\beta_{i,j,k}E_{j}E_{k}$$

where in the second term, $A_{i,jk}$ is the dipole-quadrupole polarizability term. This equation is introduced for the case of polar molecule in condensed phase and in the related paper it is for water.

It thus appears that the contribution of dipole-quadrupole polarizability is rather small but maybe higher than hyper-polarizability. 

Could someone explain or refer some literature so as to describe:

 - The importance of the cross-coupled terms like dipole-quadrupole polarizability, 
 - And, are there some cases where the cross-coupled terms become zero (and vice-versa).