“…Therefore, we may assume 1 ∈ 2V (1, 0) as the identity 1 generates a copy of V (1, 0). Thus we may choose a linear basis {x 1 , x 2 , x 3 , 1} for A such that {x 1 , x 2 } is a canonical basis for M 1 (1, 0, ∞) with the D 4 -action given by (5) and 2V (1, 0) 4 . Replacing x 3 with x 3 + α for some suitable α ∈ k, we may assume…”