As printed, the blocks split on the parity of $\{b,d\}$ — BD, not a listed effect

For a $2^4$ factorial design with four treatments A, B, C, D each at two levels, the treatment combinations were allotted in two blocks of 8 plots each as below:

$$\begin{array}{l|cccccccc} \text{Block I} & 1 & a & c & ac & bd & abd & bcd & abcd\\ \text{Block II} & b & ab & bc & abc & d & ad & cd & acd \end{array}$$

Hence the treatment combination which is confounded is: